Presenter: Prof. Dr.-Ing. João Paulo Carvalho Lustosa da Costa University of Brasília

Abstract: Nowadays the amount of subscribers connected to mobile networks exceeds three billions, and according to the Wireless World Research Forum (WWRF) it is expected to have seven billions subscribers for 2017. Moreover, high data rate applications are foreseen for the next generation (3D) mobile Internet with 3D graphics and animations. As a consequence, the wireless communication systems should evolve in order to fit to the future demands. MIMO systems show to be a promising solution for the future wireless communications systems.

MIMO systems are basically communication systems composed of multiple antennas at the transmitter and multiple antennas at the receiver together with advanced signal processing algorithms. This whole set of hardware and software increases the spectral efficiency with a low cost compared to other solutions.

In order to carry out deployment planning, algorithm design, and system simulations, a profound knowledge of the underlying physical MIMO channel is necessary. In the literature, the most common model for the physical MIMO channel is called multipath components (MPCs), where each MPC describes a discrete link between the stations characterized by physical parameters, like path loss, phase, delay, direction of departure (DoD), and direction of arrival (DoA). These unique links between DoAs and DoDs are referred as double-directional radio channel. For instance, the Wireless World Initiative New Radio (WINNER) channel model (WIM) is based on the MPCs concept. Moreover, the International Telecommunication Union (ITU) has adopted a slightly modified WINNER channel model for evaluation of IMT Advanced radio interface IMT-Advanced technologies. To provide more realistic channel models, the physical ovide parameters should be extracted from measurements given a known scenario. Therefore, the development of robust algorithms to extract these parameters of the MIMO channel is crucial.

In Fig. 1, we exemplify a MIMO channel with four MPCs. From each MPC, we can extract the following parameters: direction departure (DOD), direction-of-departure direction-of-arrival (DOA), time delay of arrival (TDOA), and Doppler shift.

Figure 1. Example of MIMO channel

In order to extract parameters, we show a parameter estimation chart in Fig. 2 with three steps: the model order selection, the parameter estimation and the subspace prewhitening. The model order selection is the estimation of the number of MPCs. The parameter estimation is the extraction of the desired parameters once the model order is estimated. Finally, the subspace prewhitening is a step that can be applied for the case that the noise is colored.

Figure 2. Parameter estimation chart

In this tutorial, we focus on presenting advanced signal processing algorithms to estimate the model order, also known as number of MPCs, as well as the associated spatial frequencies applying multi-dimensional array signal processing and taking into account the colored noise.

For each of step in Fig. 2, there are numerous schemes in the literature. Therefore, we present the state-of-the-art schemes based on multilinear algebra. Also during the tutorial, concepts of tensor calculus are reviewed.

Short bio of the presenter:

João Paulo Carvalho Lustosa da Costa received the Diploma degree in electronic engineering in 2003 from the Military Institute of Engineering (IME) in Rio de Janeiro, Brazil, his M.S. degree in 2006 from University of Brasília (UnB) in Brazil, and his Doktor Doktor-Ingenieur (Ph.D.) degree with Magna cum Laude in 2010 from Magna Ilmenau University of Technology (TU Ilmenau) in Germany.

Currently, he is a professor at the Electrical Engineering Department, University of Brasília (UnB), and he cooperates with the Laboratory of Technologies for Decision Making (LATITUDE) supported by DELL Making computers of Brazil, with the Laboratory of Automation and Robotics (LARA), and with the Microwave and Wireless Systems Laboratory (MWSL). He coordinates the Laboratory of Array Signal Processing (LASP) at UnB. His research interests are in the areas of multi dimensional array signal processing, model order selection, principal component analysis, MIMO communications systems, multilinear algebra and parameter estimation schemes.