Marko Beko is a researcher at the Institute for Systems and Robotics and a professor at the ULHT . His research work has been on coding-modulation and information-theoretic aspects of noncoherent multi-input multi-output (MIMO) communications.  In noncoherent communications, channel state information (CSI) is assumed to be unknown to both the transmitter as well as the receiver. Research in this area is considered difficult due to the absence of an explicit closed form expressions for the probability of error and mutual information at a general signal-to-noise ratio (SNR).

Contrary to other approaches, in this research, the Gaussian observation noise may have an arbitrary correlation structure. The main problems in noncoherent communications that have beeen considered are:

1. How to design MIMO signal constellation matrices  that  minimize the probability of error at high SNR for the case where the channel matrix is modeled as an unknown deterministic parameter ?

2. What are the optimal MIMO signal constellation matrices that maximize the mutual information of the spatially correlated Rayleigh fading channel at low SNR ?

3. How to design MIMO signal constellation matrices  that  minimize the probability of error at low SNR for the case where the channel matrix is modeled as an unknown deterministic parameter ?

The publications provide a more detailed description of his research.

Keywords: MIMO, noncoherent communications, convex optimization, colored noise, geodesic descent algorithm, equiangular tight frames.

Note: As a very nice by-product of this research, some new packings in the complex projective space have been constructed, and the only known results were improved. The best known packings for the real case were given by Sloane. If you have better packings, I would like to hear about it.