Seminars

The following seminars are scheduled for 2007/2008:

1 Speaker: Manuel de Oliveira Duarte
Title: Evolution trends in the telecommunications sector
30.11.2007, 14h00
2 Speaker: Armando Nolasco Pinto, UA/DETI
Title: Quantum Optical Communication Systems
30.11.2007, 16h00
3 Speaker: Nuno Borges Carvalho, UA/DETI
Title: RF solutions for Indoor Location Based Systems
07.12.2007, 14h00
4 Speaker: José Carlos Pedro, UA/DETI
Title: Digital Transceiver Architectures – A New Paradigm in Wireless RF Sub-System Design
07.12.2007, 16h00
5 Speaker: Carlos Lima, DEI/UM
Title: Smart Antennas
14.12.2007, 14h00
6 Speaker: João Barros, FCUP
Title: Physical-layer Security: From Theory to Practice
14.12.2007, 16h00
7 Speaker: Luis Almeida, UA/DETI
Title: Real-time traffic scheduling issues in distributed embedded systems
01.02.2008, 14h00
8 Speaker: José Afonso, DEI/UM
Title: Quality of Service in Wireless Sensor and Actuator Networks
01.02.2008, 16h00
9 Speaker: Filipe Sousa, INESC Porto
Title: Network Monitoring
08.02.2008, 14h00
10 Speaker: Susana Sargento, DETI/UA
Title: Integration of Heterogeneous Networks
08.02.2008, 16h00
11 Speaker: Joaquim Neves, DEI/UM
Title: Global Information Infrastructure/Next Generation Network
15.02.2008, 14h00
12 Speaker: Rui Campos, INESC Porto
Title: Spanning Trees in Communication Networks
15.02.2008, 16h00
13 Speaker: José Vieira, UA/DETI
Title: Real Number Coding and Compressed Sensing
22.02.2008, 14h00
14 Speaker: Paulo Jorge Ferreira, UA/DETI
Title: Digital Fountains
22.02.2008, 16h00
15 Speaker: A. Gameiro, UA/DETI
Title: Spectrum Efficiency and Opportunistic / Cognitive Radio Communications
29.02.2008, 14h00
16 Speaker: Miguel Rodrigues, FCUP
Title: Resource Optimization in Communications Systems: An Information Theoretic-Estimation Theoretic Approach
29.02.2008, 16h00
17 Speaker: José Rocha Pereira
Title: Computational electromagnetics in time domain
07.03.2008, 14h00
18 Speaker: Rui L. Aguiar, UA/DETI
Title: On clean slate design
07.03.2008, 16h00

.

Seminar 1

Speaker: Manuel de Oliveira Duarte
Title: Evolution trends in the telecommunications sector
Abstract:
This seminar will present an overview of some of the most significant developments in the telecommunications sectors over the last decades and will attempt to identify some of its possible future evolutionary paths in terms of technology and regulation. In more detail the seminar will address such issues as follows:

  • Transformations in the telecommunications sector have unfolded in a vertiginous manner over the last decades, in a succession of interrelated processes that is continuing with unabated vigour.
    On the technological side the following drivers deserve special mention:
    • Continuing progresses in microelectronics miniaturisation and large-scale integration, enabling increased processing capabilities and computing power, leading to things that think.
    • Continuing massive increases in communications power and network capacity, leading to things that link.
    • Widespread digitisation of transmission and switching mechanisms, the ever-increasing utilisation of optical fibres as transmission media, and the introduction of sophisticated software control in almost all parts of networks.
    • Synergies between packet switching networks and hypertext based applications leading to the Internet explosion.
    • Emergence of the concept of web based network computing as the dominant environment for cooperative and distributed applications and services.
    • A move away from the dominance of the keyboard and screen to such interfaces as voice, speech and pattern recognition in voice-webs, 3-dimensional displays, and automatic data capture (especially through various forms of sensors).
    • A move towards integrated hand-held devices, which combine two or more functions (e.g. computing, telephone functions, Internet access, etc.), such as PDAs, palm devices, the “office-in-the-pocket”, etc.
    • Introduction of wireless technologies (GSM, GPRS, UMTS, UMTS, Hiperlan2, Bluetooth, etc) which may replace traditional one-design-fits-all services with individualised systems which respond to individual terminals or agents triggered by smart profiling.
    • Convergence and integration of technologies (e.g. the PC and the telephone, Internet and Television Services, fixed and mobile communications); computing components are becoming embedded into non-desktop devices which are thus becoming smaller, lighter and interconnected.
    • New materials are becoming available and used, for example, in flat displays, smart textiles and even smart ink.
    • Developments in artificial intelligence, which assist the user in finding, sifting, storing and retrieving information. At the present time, these include data warehousing and data mining processes, but strong developments are also taking place in software agents (which can be programmed to reflect user preferences) and context-sensitive knowledge management tools.
  • Market and regulatory forces have driven major industry transformations:
    • From a heavily regulated, “end-to-end” and “natural monopoly”, telecomms have become...
    • ... an industry with segments separated by regulation and transformed by selected competitive entry, in a largely liberalised market, moving toward…
    • ...a single super-industry with few intra-industry barriers to entry: cable, wireless and internet-based technologies enabling entry by start-ups, industry incumbents and firms from other industries thought previously to be “unrelated.”

In socio-economic terms the predicted continuing growth of the Internet, mobile phone services and other telematic applications is penetrating all forms of life, at the individual and at the collective level, starting to create a networked society, in which large sectors of the population will expect to be able to communicate anywhere, anytime and using any medium that is always online. The need here is for seamless interoperability where devices and services become platform independent. This aim faces several challenges, amongst them lack of standards and uncertainty regarding market regulation, proprietary ownership and intellectual assets which policy needs to address.
Another major challenge is the risk of what is commonly called the digital divide: the fracture between those who can afford and have the ability to take advantage of the more advanced technologies and services and those who either cannot afford or do not possess the adequate technical literacy to use those resources, or live in areas where they are simply not available.
In this way, technologies and telecommunications services also become part of the struggle for better standards of life of many citizens throughout the world. Excellent progress has been made over the last ten years but many problems remain where availability and affordability are not equally developed across the whole of the EU, or between different market (user) segments and groups. Many bottlenecks are caused by large geographic and group variations in skills and awareness (due to the enormous learning costs involved), so that markets fail to function optimally and thus do not achieve the growth and returns they should, nor do they fully contribute to economic growth, competitiveness and employment. For example, despite general regional convergence in socio-economic measures like GDP per capita in Europe over the last ten years, there are increasing regional disparities in terms of technology availability. The long, and on-going, discussion about the USO (Universal Service Obligation), which seems to have become buried over the last few years, is currently being revived.
Faced with the above trends and challenges, this Seminar will promote a exchange of ideas about possible future evolutionary paths in terms of technology and regulation, capable of making telecommunications technologies and services more effective and more reachable for the common citizen.

Seminar 2

Speaker: Armando Nolasco Pinto, UA/DETI
Title: Quantum Optical Communication Systems
Abstract:
During the last fourteen years, I have been working in several aspects related with the analysis, modeling, design and optimization of optical communication systems. During this period the capacity of the systems have increase by more than three orders of magnitude, from less than 1 Gbps to more than 1 Tbps. However, the demand for bandwidth has increased even at a higher ratio. The huge demand of bandwidth is greatly justified by the surge and democratization of the Internet and mobile networks.
It is my belief that if we want to increase substantially the systems capacity we have to model the systems with greater accuracy. Traditionally, optical communication systems have been modeled using the classical electromagnetic theory. A deeper understanding of the systems can only be achieved if we adopt a quantum description of the involved processes.
At a fundamental level, all physical communication channels are subject to the laws of quantum mechanics. But while quantum effects are negligible for radio-frequency systems, they can be a dominant factor at optical frequencies. Electrical communication engineers worry about noise because they have to discern weak signals in the presence of background noise. At frequency below the far-infrared, shot noise and thermal noise dominates, at higher frequencies quantum noise is predominant. Therefore in order to push further the capacity of optical communication systems it is necessary to understand quantum noise.
The universe offers many more ways to communicate and manipulate information than we presently use. The laws of quantum mechanics not only place limitations, they also offer resources for enhancing our ability to communicate. Quantum mechanics predicts that strong correlations known as entanglement can exist between separate quantum systems. Nowadays, there are already commercial systems using quantum entanglement to achieve secure communications. In the future, other disruptive technologies will appear and it is my strong belief that quantum theory will play a key rule in this process.
In this talk, I will present a short overview of the quantum theory, followed by several examples where quantum theory can be used to gain a deeper understanding of classical optical communication systems or in order to create system with new functionalities only achieved by quantum manipulation. I will present cutting-edge research work done within my research group and within other research groups in this fascinate field.

Seminar 3

Speaker: Nuno Borges Carvalho, UA/DETI
Title: RF solutions for Indoor Location Based Systems
Abstract:
This talk is aimed at the explanation of location and positioning technologies for indoor environments. In the seminar some techniques for location systems, as angle of arrival, time of flight, differential time of flight, and RF fingerprinting, will be addressed and explained. It is expected to be able to make an overview of the most used schemes for this type of systems, and present some practical solutions for location in indoor environments.
The main problems associated with these schemes will also be addressed as some proposals for error minimization, in the path for this explanations basic concepts as system modeling using nonlinear approximants will also be addressed, specially neural networks modeling approaches for location schemes.

Seminar 4

Speaker: José Carlos Pedro, UA/DETI
Title: Digital Transceiver Architectures – A New Paradigm in Wireless RF Sub-System Design
Abstract:
The possibility of incorporating digital signal processing techniques in current wireless transceivers, has opened new fields of research in the RF sub-system architecture, promising a shift of paradigm on near future wireless circuit design. So, this Seminar aims at describing the advances and challenges that are expected with such a change of paradigm.
For that, the talk will address both the system architecture and RF circuit implementation levels. In the first case, we will briefly describe the newly proposed digital LINC, Cartesian and Polar TXs, and compare them to the conventional analog Cartesian or I/Q modulator based architecture. Although theory has predicted outstanding performances from these digital architectures, their conversion into real wireless RF front-ends is limited by the non-idealities of RF circuitry. Therefore, we will show how it is needed to put together researchers of RF electronics, control theory and power supply converter technology to address both the system level and implementation aspects in an integrated way.

Seminar 5

Speaker: Carlos Lima, DEI/UM
Title: Smart Antennas
Abstract:
A smart antenna system combines multiple antenna elements with a signal-processing capability to optimize its radiation and/or reception pattern automatically in response to the signal environment.
A wide range of wireless communication systems may benefit from spatial processing, including high-mobility cellular systems, low-mobility short-range systems, wireless local loop applications, satellite communications, and wireless LAN. By employing an array of antennas, it is possible to multiplex channels in the spatial dimension just as in the frequency and time dimensions. To increase system capacity, spatially selective transmission as well as spatially selective reception must be achieved.
Adaptive arrays utilize sophisticated signal-processing algorithms to continuously distinguish between desired signals, multipath, and interfering signals as well as calculate their directions of arrival. This approach continuously updates its transmit strategy based on changes in both the desired and interfering signal locations. The ability to track users smoothly with main lobes and interferers with nulls ensures that the link budget is constantly maximized because there are neither microsectors nor predefined patterns.
The flexibility of adaptive smart antenna technology allows for the creation of new value-added products and services that give operators a significant competitive advantage. Adaptive smart antennas are not restricted to any particular modulation format or air-interface protocol. They are compatible with all current air-interface modulation schemes.
The main purpose of this tutorial is to present the most relevant algorithms in array signal processing and the current research directions, especially concerning to DoA algorithms for arbitrary geometry arrays.

Seminar 6

Speaker: João Barros, FCUP
Title: Physical-layer Security: From Theory to Practice
Abstract:
Recent theoretical and practical work has shown that novel physical layer security techniques have the potential to significantly strengthen the security of wireless networks. In the first part of this talk we will briefly review the fundamentals in information-theoretic security and discuss our most recent results.
Formulating the problem as one in which two legitimate partners communicate over a quasistatic fading channel and an eavesdropper observes their transmissions through another independent quasi-static fading channel, we define the secrecy capacity in terms of outage probability and provide a complete characterization of the maximum transmission rate at which the eavesdropper is unable to decode any information. In sharp contrast with known results for Gaussian wiretap channels (without feedback), our results show that in the presence of fading information-theoretic security is achievable even when the eavesdropper has a better average signal-to-noise ratio
(SNR) than the legitimate receiver. The role of channel state information shall be discussed in detail.
Another relevant line of research addressed in this talk is directed towards understanding the security implications of the emerging network coding paradigm. Here, intermediate nodes in the network are no longer limited to storing and forwarding packets but instead are allowed to mix different data flows using linear operations, thus achieving higher throughput and robustness among other benefits. We shall show that random linear network coding provides intrinsic algebraic security against both external wiretapping and internal eavesdropping attacks, as well as Byzantine network pollution. A rigorous characterization of the fundamental security shall also be included in the discussion.
In the second part of the talk, we shall show how these theoretical results can be translated into usable physical-layer security technologies.
We start by presenting a practical security scheme by which two terminals (say Alice and Bob) are able to exploit the randomness of wireless fading channels to exchange data in an information-theoretically secure way. To ensure that a potential eavesdropper (say Eve) is unable to decode any useful information, Alice sends useful symbols to Bob only when the instantaneous secrecy capacity is strictly positive. In the remaining time, a specially designed class of LDPC codes is used for reconciliation, thus allowing the extraction of a secret key, which can be distilled using privacy amplification. We believe this opportunistic approach can be used effectively as a physical layer complement to existing cryptographic protocols. Another practical example builds on the exploitation of the intrinsic security provided by network coding to develop secure broadcast protocols for wireless networks.
Joint work with Miguel Rodrigues (FCUP), Luisa Lima (FCUP), Muriel Médard (MIT), Matthieu Bloch and Steve McLaughlin (Georgia Tech).
For more information: http://www.dcc.fc.up.pt/~barros

Seminar 7

Speaker: Luis Almeida, UA/DETI
Title: Real-time traffic scheduling issues in distributed embedded systems
Abstract:
Many distributed embedded systems are real-time systems in which responses to input events must be generated within bounded time windows. This end-to-end constraint reflects itself at all levels of the system and, particularly, at the network level, where appropriate protocols with bounded latencies must be used. In this talk we will briefly survey the most common protocols used within the real-time distributed embedded systems domain, e.g., CAN, WorldFIP, TTP/C, FlexRay, and Ethernet variants, focusing on paradigmatic real-time scheduling techniques to determine upper bounds to the network-induced delay. Then, we will focus on the use of temporal partitions over such networks, and on the techniques available to analyze timeliness within the partitions. Finally, we will see a couple of anomalies that can occur with such framework, related with the definition of a critical instant. The talk highlights such anomalies showing their origin and indicating that it is not possible, for those situations, to determine the worst-case network delay with exactitude in the general case. Finally, the talk ends with several open issues in traffic scheduling for real-time networks.

Seminar 8

Speaker: José Afonso, DEI/UM
Title: Quality of Service in Wireless Sensor and Actuator Networks
Abstract:
The field of wireless sensor and actuator networks (WSANs) has advanced considerably over the last years and new research problems are emerging. In particular the aspect of quality of service (QoS) is becoming more important, driven by applications such as distributed control systems and real-time monitoring systems. These applications present QoS requirements such as sustained throughput, bounded delay and packet delivery reliability, which are hard to achieve in the context of wireless sensor and actuator networks, due to some common attributes expected from these networks, such as small energy budgets, low duty cycles, node failures and multihop communications. While the topic of quality of service has received a lot of attention in other wired and wireless communications systems, there are still few contributions in the literature dedicated to the field of wireless sensor and actuator networks.
This presentation will provide an overview of the topic of quality of service in wireless sensor and actuator networks, including a discussion of potential applications, challenges and research directions. It will also describe the research and development activity in this area carried out at the Department of Industrial Electronics of University of Minho, with focus on the systems and solutions developed under the scope of two funded R&D projects:

  • A real-time wireless sensor network of multiple sensor-equipped garments for hidrocinesiotheraphy.
  • A fly-by-wireless onboard wireless sensor and actuator network for unmanned aerial vehicles (UAVs).

Seminar 9

Speaker: Filipe Sousa, INESC Porto
Title
: Network Monitoring
Abstract:
Network monitoring is currently considered one of the main study areas in network management. Networks are growing very fast and users demand for new services (e.g. VoIP and VideoConference) with Quality of Service. Moreover, new business models demand for complex tariffs based on usage-based accounting or hybrid ones (mixture of event based and time based). Security is also becoming the main concern for users, regarding aspects as confidentiality, privacy or authenticity. We can consider five main driver applications for network monitoring:

  • Monitoring Usage-based accounting is one of the target applications required by usage-based billing-systems. In accordance to the tariff model we need to perform accounting per flow.
  • Traffic profiling, measurements captured over a long period of time can be used to track and anticipate network growth and usage. Such Information is valuable for trend analysis and network planning.
  • Traffic engineering, aims at the optimization of network resource utilization and traffic performance.
  • Attack and intrusion detection are among the main target applications for network monitoring. The number of metrics useful for attack detection is as diverse as attack patterns themselves. Attackers adapt rapidly to circumvent detection methods and try to hide attack patterns using slow or stealth attacks.
  • QoS monitoring is the passive observation of the transmission quality for single flows or traffic aggregates in the network. One example of its use is the verification of the conformance of the QoS guarantees in service level agreements (SLAs).

The main challenges for this subject are how to perform the measurements for all these applications without been very intrusive i.e. injecting a lot of packets into to networking in order to get the required metrics. Another important issue is the information entropy; the idea is to have reliable measurements results with less information and also reuse the same sample for several applications. This can be achieved by adopting sampling techniques, i.e. instead of capturing all the information we only capture a subset that assures reliable measurements.

Seminar 10

Speaker: Susana Sargento, DETI/UA
Title: Integration of Heterogeneous Networks
Abstract:
Objectives: The goal of this seminar is to provide an overview of cutting-edge research on the integration of heterogeneous networks, where users can seamlessly move between any of the available networks. This class will address the main issues of mobility and QoS in a heterogeneous environment, as well as the support of the integration of new networks, such as ad-hoc, mesh and moving networks. It offers the students the opportunity to grasp the fundamental principles of heterogeneous networking and to understand the main problems that arise in an integrated environment.
Summary: With the current evolution of network technologies we envision that, in a near future, there will be a heterogeneous landscape of different technologies such as WLAN, WiMAX, UMTS/MBMS and DVB, providing ubiquitous network access to users. Also, networks such as ad-hoc and moving networks are also expected to be integrated in a next generation network. Regarding network providers, multitechnology and multi-network availability can and should be used to enhance user experience aiming at the always-best-connected paradigm.
This ubiquitous network access requires an architecture that homogeneously integrates the heterogeneous access technologies for seamless movement of the users throughout the available technologies.
This seminar will comprise a specific range of topics with a focus on the integration of heterogeneous networks, covering important research results on mobility, multihoming, QoS from the past and promising new directions. Topics that will be covered include:

  • Heterogeneity of access technologies: local area technologies, cellular technologies
    and broadcast distribution technologies
  • Integration of heterogeneous technologies: IP(v6) as the support for integration
  • Challenges for networks integration:
  • Mobility: new fast mobility approaches and media independent, multihoming
  • QoS: architectures for QoS support (efficiency and scalability) and
    integration with mobility
  • Support for the integration of ad-hoc, mesh and moving networks

Seminar 11

Speaker: Joaquim Neves, DEI/UM
Title: Global Information Infrastructure/Next Generation Network
Abstract:
By definition, the Global Information Infrastructure (GII) will be an infrastructure which facilitates the development, implementation and interoperability of existing and future information services and applications within and across the telecommunications, information technology, consumer electronics and content provision industries. This infrastructure will consist of interactive, broadcast and other multimedia delivery mechanisms coupled with capabilities for individuals to securely share, use and manage information, anytime and anywhere, with security and privacy protection, and with levels of acceptable cost and quality.
Taking into consideration the extraordinary expansion of digital traffic, and the need to converge and optimise the operating networks, several study groups are developing global standards for Next Generation Network (NGN) at the International Telecommunication Union (ITU ), which will establish the convergence of telecommunications network and IP network technologies. By other side, TISPAN (Telecommunication and Internet converged Services and Protocols for Advanced Networking) is the European Telecommunications Standards Institute (ETSI) core competence centre for fixed networks and for migration from switched circuit networks to packet-based networks with an architecture that can serve in both to create the Next Generation Network.
A Next Generation Network (NGN) is a packet-based network able to provide Telecommunication Services to users and able to make use of multiple broadband, QoS-enabled transport technologies and in which service-related functions are independent of the underlying transport-related technologies. It enables unfettered access for users to networks and to competing service providers and services of their choice. It supports generalized mobility which will allow consistent and ubiquitous provision of services to users.
The proposed 2 hours session of the MAP-Tele Seminar Curricular Unit, will focus on the recent activities of the Next Generation Network - Global Standards Initiative (NGN-GSI), that are being held at ITU.

Seminar 12

Speaker: Rui Campos, INESC Porto
Title: Spanning Trees in Communication Networks
Abstract:
State of the art routing protocols find the best unicast routes within a given communication network based on Shortest Path spanning Trees (SPTs), computed from each network node perspective; the union of these trees defines the set of active network links. Minimum Spanning Trees (MSTs) have been used in finding the optimal multicast/broadcast tree, whereas the Minimum Routing Cost spanning Tree (MRCT) defines the optimal solution when the active network topology has to be a tree. These just represent some examples of the extensive use of spanning trees in the scope of communication networks.
In this talk we will start by introducing the concepts of graph and spanning tree. Afterwards, we will go through the traditional algorithms from graph theory used for finding SPTs (e.g., Dijkstra, Bellman-Ford) and MSTs (e.g., Kruskal, Prim). In addition, we will present some approximate algorithms currently used to compute the approximate MRCT, given that its exact computation is known to be an NP-hard (Nondeterministic Polynomial-time hard) problem. More specific algorithms considering the combination of SPT and MST algorithms will also be presented taking into account specific applications in communication networks. In the end, we will show some recent research work regarding the properties of SPTs when the whole spectrum of communication networks, from homogeneous to highly heterogeneous networks, is considered. We will show, for instance, that concerning highly heterogeneous networks, the union of SPTs coincides with the MST. On the other hand, we will characterize the differences of using single and multiple spanning trees in communication networks, namely the gains and the losses.

Seminar 13

Speaker: José Vieira, UA/DETI
Title: Real Number Coding and Compressed Sensing
Abstract:
Digital coding is usually performed in finite fields as the GF(28). The algebra of finite fields is counter intuitive and as a consequence for most of the students it is difficult to see the principles behind the algorithms. Blahut and Marshall were one of the first researchers to notice the tight connections between the signal processing and the coding areas. Coding using the real or complex fields was until a few years a mere academic topic with only a pedagogical value.
Recently, Elad and Donoho discover a method to recover a sparse signal from random linear combinations of a signal. This result can be applied to the finite and real number fields.
In this talk we will give a signal processing perspective of some well known codes with the purpose to enlighten the mentioned connections. We also will show the importance of the new random codes and their utilization as Digital Fountains.

Seminar 14

Speaker: Paulo Jorge Ferreira, UA/DETI
Title: Digital Fountains
Abstract:
Digital Fountain codes is a new concept where the traditional paradigm of transmitting information as an ordered stream of packets is changed to one where the user must receive a sufficient amount of packets in order to reconstruct the original information.
In this talk we will start with a brief description of the evolution of this type of codes comparing their performance with the traditional codes. An overview of the applications of Digital Fountains will also be presented.

Seminar 15

Speaker: A. Gameiro, UA/DETI
Title: Spectrum Efficiency and Opportunistic / Cognitive Radio Communications
Abstract:
To present the motivation new spectrum management models, identify the options at stake and present the main challenges related to the implementation of opportunistic radio systems.
Outline:

  • The spectrum problem
  • Spectrum scarcity: myth or reality
  • The GOSPLAN method of spectrum management
  • The new models for spectrum management
  • Spectrum to the commins
  • Spectrum to the market
  • Fee simple ownership with non-interference easement
  • Frequency agile systems
  • The concept
  • The challenges
  • Architectures – The main options
  • Inter-operator cooperative management
  • Sensing based management
  • The sensing component
  • Feature detectors
  • Cooperative sensing
  • Resource management
  • Competition vs coopreation
  • Basics of game theory and its application in opportunistic radios

Seminar 16

Speaker: Miguel Rodrigues, FCUP
Title: Resource Optimization in Communications Systems: An Information Theoretic-Estimation Theoretic Approach
Abstract:
A large body of optimization problems abounds in the telecommunications field, particularly enticing ones relating to resource optimization in communication systems. In this talk, we will concentrate on power allocation optimization problems in so-called multiple input-multiple output Gaussian channels with arbitrary inputs, that represent a very large number of relevant communications scenarios such as multiple antenna systems, code division multiple access (CDMA) systems or digital subscriber line (DSL) systems. In this context, we will resort to a key information theoretic-estimation theoretic mathematical framework to tackle this kind of problems in an effective manner.
Initially, we will review some basic principles of unconstrained and constrained optimization theory, with a particular emphasis on convex optimization problems. We will introduce a number of key concepts including the first-order or Karush-Kuhn-Tucker (KKT) optimality conditions that yield the various critical points (maxima, minima and saddle-points) to a particular optimization problem.
We will then consider a number of classic power allocation problems in communications systems, namely power allocation in parallel independent Gaussian channels with Gaussian inputs as well as parallel independent Gaussian channels with arbitrary inputs. In this context, we will review the conventional waterfilling and mercury-waterfilling graphical interpretations that illustrate the optimal power allocation policy that maximizes the system mutual information (or system information rate).
We will later consider power allocation in multiple input-multiple output Gaussian channels with arbitrary inputs. Here, we will exploit the information theoretic-estimation theoretic framework to determine the power allocation strategy that maximizes the information rate between the input and the output of the system. We will also put forth a novel mercury-waterfilling interpretation of the optimal power allocation procedure that generalizes the conventional waterfilling and mercury-waterfilling interpretations.
We will also exploit this same mathematical framework to specialize the optimal power allocation strategy to asymptotic regimes, namely the low- and high-power regimes. Of particular relevance, we will demonstrate that in the high-power regime the power allocation strategy that maximizes mutual information also maximizes the minimum distance of the output lattice, hence minimizes the error probability. We will also extrapolate these results to optimal precoding policies.
This represents joint work with Prof. Fernando Perez-Cruz and Prof. Sergio Verdú from the Department of Electrical Engineering, Princeton University, U.S.A.

Seminar 17

Speaker: José Rocha Pereira
Title: Computational electromagnetics in time domain
Abstract:
Computational electromagnetics is a very active area and in the last decade was a very interesting increasing for a wide class of applications. For the numerical solution of such problems, Yee's Finite-Difference Time-Domain (FDTD) scheme has become a quite popular technique. FDTD is related to the physical space representations of the fields, in which the electric and magnetic fields are described by vectors, both discretized in time and space in a grid usually uniform. The main disadvantage of this representation is therefore the possible increase of the computation time to simulate such grid. For this reason, it is important to study adaptive numerical methods that use a refined grid only in certain regions of the space, and a less refined grid in other regions where the variation of the fields is smoother.
In the wavelet context, the Multiresolution Time-Domain Technique (MRTD) scheme offers a natural framework for the definition of dynamic adaptive strategies for Maxwell equations. However, in general the implementation of this scheme is complex, which seriouslycompromises the computational performance.
There is another type of adaptive strategy named SPR (Sparse Point Representation) that uses interpolating wavelets tools. The principle of the method is to represent the solution only through those point values indicated by the significant wavelet coefficients, which are defined as interpolation errors. Typically, few points are needed in each time step, the grid being coarse in smooth regions, and refined close to irregularities.
Topics to be addressed:

  • Yee's Finite-Difference Time-Domain (FDTD) scheme.
  • Brief presentation of the Multiresolution Time-Domain Technique (MRTD) scheme.
  • SPR (Sparse Point Representation) scheme.
  • Stability and dispersion analyses.

Seminar 18

Speaker: Rui L. Aguiar, UA/DETI
Title: On clean slate design
Abstract:
Current network design is based on the Internet paradigm. The basic IP-design considerations have long been challenged by issues as mobility, QoS, multimedia, security and multicast. As a consequence, the IP deficiencies are well-known. Recently, some advocates of a refunding of networking principles have been presenting a radical change on networking design, the so-called clean-slate design. The talk will address the major challenges being tackled by these novel approaches.