University of Rostock

Several key technologies are currently discussed for the evolution of next generation mobile cellular radio systems (5G). While multiple antenna systems have already been established in current standards, different extensions have the potential to be integral part of future systems. Deploying a huge number of antennas at the base station (massive MIMO) is one proposal to overcome interference limitations in cellular networks. Moreover, spatial modulation is a technique that transmits information by selecting only one out of N antennas. Protagonists of this approach claim that it saves hardware costs and reduces power consumption by replacing multiple power amplifiers and filters by a single hardware chain and a fast switch connecting the chosen antenna to the hardware chain.

In the field of digital signal processing, compressed sensing has attracted a lot of attention in the last Spatial Modulation (SM) is a recently developed Multiple-Input Multiple-Output (MIMO) transmission technique. The basic concept of SM is that information is transmitted using two information carrying units. Information is mapped to symbols from a signal constellation and additional information is transmitted by selecting a unique transmit antenna pattern from a set of possible antenna patterns. SM has attracted academic attention due to its capability of eliminating inter channel interference and avoiding inter antenna synchronization.

Despite a large number of publications on spatial modulation, there are still several open research issues. For instance, many SM approaches cannot be easily scaled to massive MIMO scenarios. There exist low-complexity detection techniques that achieve near-optimum performance for relatively small antenna and signal constellations. However, these algorithms do not scale well for large antenna systems with high spectral efficiency.

Although error-correcting codes are applied in practically all communication systems, the majority of the publications on SM consider only uncoded transmission. This probably results from the fact that coded SM transmission requires new coding approaches like non-linear codes for protecting the antenna selection patterns. Today’s code designs based on trellis codes or space-time codes are only suitable for very small systems. Moreover, new signal constellations were developed that outperform conventional two-dimensional signal constellations for SM and need novel coding techniques as well. Up to now, such codes have found little consideration in the coding community.

The aim of this research project is the development of new coding schemes for spatial modulation. Monolithic coding approaches will be pursued for systems with just a few antennas. Larger systems require a separation of antenna selection and encoding of data symbols due to complexity reasons. Here, multi-level coding approaches will be applied to enable independent encoding of the antenna patterns and signal points. New low complexity encoding and decoding strategies for information transmission via antenna selection will be developed.

Noteworthy, the sets of antenna patterns will be considered as non-linear constant-weight codes and their concatenation with trellis or algebraic codes copes with the non-linearity and the complexity of this coding problem. Furthermore, lattice-based signal constellations proposed by the applicants have shown remarkable good performance for spatial modulation. The algebraic structure of these constellations will be exploited for novel coding schemes. Finally, new probabilistic coding techniques like Sparse Regression Codes will be applied to SM. Thus, we expect many interesting results for the improvement of future coded spatial modulation systems.