• Red TEIC
  • Red TEIC
  • Red TEIC

Encuentro de expertos

Los profesores invitados impartirán charlas magistrales: Mónica F. Bugallo (Stony Brook University, USA), Michael Joham y Andreas Gründinger   (Technical University of Munich, Germany) y Carlos Escudero  (University of A Coruña, Spain).

2 Junio 2015, 9:45 - 13:30

Área Científica (AC) Building, Room A.3.01.AC, Campus de Elviña, A Coruña.

Program

Time  Content Speaker
9:45-10:00 Welcome reception Room A.3.01AC
10:00-11:00 Monte Carlo Methods for Complex Systems

Mónica F. Bugallo, Stony Brook University, USA

11:00-12:00 Strategies to Combat Pilot Contamination in   Massive MIMO Systems Michael Joham, Technical University of Munich, Germany
12:00 -12:30

Outage Constrained Downlink Beamforming with   Multiplicative and Additive Channel Errors

Andreas Gründinger, Technical University of Munich, Germany
12:30-13:00 The GTEC Unmanned Aerial Vehicle (UAV): an indoor navigation demonstration Carlos Escudero, University of A Coruña, Spain
13:00-13:15 Conclusions Luis Castedo, University of A Coruña, Spain
13:30 Lunch Facultad de Informática


Meeting hosting

Luis Castedo
University of A Coruña, Department of Electronics and Systems
Facultad de Informática, Campus de Elviña s/n
15071, A Coruña, Spain

Keynote talks

Monte Carlo Methods for Complex Systems

Mónica F. Bugallo

Department of Electrical and Computer Engineering, Stony Brook University, USA

 

Advances in the development of models and methods that can satisfactorily describe and analyze high dimensional systems are extremely valuable for many different disciplines including biology, meteorology, economics, social sciences, and engineering. These systems are characterized by nonlinearities and are difficult to understand. Computational methods governed by simple local rules have the potential of providing insightful interpretations and of paving the way towards quantitative and qualitative descriptions and understanding of complex systems.

This talk discusses the application of the sequential Monte Carlo methodology to high dimensional problems.

In the literature, there are claims stating that particle filters cannot be used for complex systems because their random measures degenerate to single particles. While this is true for standard implementation of these filters, it does not hold true for alternative approaches. New alternatives based on the principle of divide and conquer and on marginalization of nuisance parameters are presented. In particular, the collapse of traditional particle filtering is avoided by setting an interconnected network of filters, each of them working on lower dimensional spaces. Current work addresses the development of theoretical grounds of the methods, establishment of guidelines for their use by practitioners, analysis of their accuracy, stability and scalability, and their validation on a wide range of complex systems. The proposed methods are original and provide solutions to arguably the biggest open problem of particle filtering.

 

Strategies to Combat Pilot Contamination in Massive MIMO Systems

Michael Joham

Signal Processing Associate Institute, Technical University of Munich, Germany

 

Massive Multiple-Input Multiple-Output (MIMO) or large-scale MIMO is foreseen to be one of the key elements of future fifth generation wireless communication systems. In massive MIMO cellular systems, the mobiles are served by base stations with a very large number of antennas (about two orders larger than the number of users). Due to the resulting quasi-orthogonality of the channels, simple operations like matched filtering are well suited to separate the signals of the different users and, theoretically, unlimited data rates are posible in massive MIMO systems. However, such a conclusion is only possible for the case of perfect channel state information (CSI) at the base stations. Interference from neighboring cells during channel estimation, so called pilot contamination, destroys the orthogonality of the channels and leads to a limitation of the possible data rates.

This talk discusses three strategies to reduce or, theoretically, eliminate pilot contamination based on channel distribution information (CDI). First, the potential of reducing the contamination effect by coordinating the pilot sequence allocation in different cells of the cellular massive MIMO system is highlighted. Second, a maximum a-posteriori (MAP) based semi-blind cannel estimation for massive MIMO systems is introduced and compared to existing pilot-assisted and blind approaches. Finally, a generalization of pilot contamination precoding is proposed that is capable to eliminate the effect of pilot contamination by jointly processing the signals of different base stations in the cellular massive MIMO system.

 

Outage Constrained Downlink Beamforming with Multiplicative and Additive Channel Errors

Andreas Gründinger

Signal Processing Associate Institute, Technical University of Munich, Germany

 

Rate balancing with outage constraints was recently considered for the beamformer design with only imperfect channel knowledge at the transmitter. We study the the multi-user vector downlink in this context. This setup is especially difficult due to the correlations of the intended useful signal power and the experienced interference.

Motivated by satellite communications, the transmitter’s cannel model contains an additive and a multiplicative random error due to multi-path scattering and shadow fading, respectively. We split the outage probabilities into two parts for the two channel errors. For a given split, the worst-case multiplicative errors are found and decent outage constraint approximations can be used for the optimization additive errors, e.g., an uncertainty reformulation or a Bernstein type bound. Finding the optimal split for the two error types resembles the maximization of a lower bound for the achievable rates. We compare an equal design with a locally optimal iterative search for this task.

Numerical results show that the equal design results meet that of the more complex iterative search if the shadow fading is similar for the users. However, the obtained performance can still be improved via a a postprocessing power allocation for example.

 

The GTEC Unmanned Aerial Vehicle (UAV): an indoor navigation demonstration

Carlos J. Escudero

Grupo de Tecnología Electrónica y Comunicaciones (GTEC), University of A Coruña

 

Robust velocity and position estimation at high rates is a critical proccess for UAVs (Unmanned Aerial Vehicles), and it is an open issue, specially indoors, where there is not GPS coverage and the environment is very aggressive. Combination of several sensors is neccesary in order to provide a reliable location system avoiding the manipulation of a human pilot (where the law permits), opening a huge ammount of real applications.

The presentation will end with a real-time demonstration to show initial advances achieved at the GTEC Lab with its own quadcopter equipped with different kind of sensors (IMU sensors, Ultrasounds, LIDAR, UWB, artificial visión). The GTEC UAV is a perfect setup for the assessment of new algorithms in a real indoor 3D environment for the real-time testing of any kind of communications, navigation, artificial visión and control system.

 

Agenda del encuentro de expertos (pdf).