Technical Program SSIE 2006

Monday June 26, 2006

9:00-12:30, Tutorial Session
Prof. Giorgio Franceschetti, University Federico II of Napoli, Italy and UCLA, USA, ‘An overall vision of electromagnetic propagation in built-up areas
Abstract: Propagation in the urban environment is in principle a well-posed problem: the electromagnetic field radiated by prescribed sources must fulfil Maxwell equations in the open space and boundary conditions over the built-up buildings walls. But any attempt to provide a solution to this apparently simple problem must face the complication of the scattering scenario. Each city is different from any other city, type and distribution of buildings may drastically change among different sections of the same city, and the electromagnetic boundary conditions may change in time, because the city is a living organism, with its cycles and temporal variations. A general frame to locate above mentioned problem is certainly desirable, if not necessary. The cities’ distribution may be considered as a stochastic process, each city being an element of this ensemble. Accordingly, the study of electromagnetic propagation and scattering in the city becomes the problem of searching the solution of Maxwell equations in a stochastic environment. This task may be pursued on along essentially two lines of thought: either a deterministic or a stochastic one. The former makes reference to an element of the ensemble, the latter exploits the statistical properties of the distribution. In the first case, the Deterministic Geometrical Model (DGM), an element of the cities’ ensemble is chosen, namely the particular city of interest. Knowledge of the three-dimensional geometry of the city must be known, i.e., shape and location of each building. These are schematised in terms of a parallepipedic structures, and ray tracing procedures are implemented to compute the electromagnetic field everywhere outside, i.e., in the streets and squares of the city, and perhaps also inside the buildings. In the second case, the Stochastic Environment Model (SEM), a totally different philosophy is followed: aim of the model is to derive general analytical expressions, describing the average properties of the urban propagation. In addition, the analytical results are required to containing a minimum number of physically meaningful parameters. An interesting model for the cities’ stochastic distribution is the percolative lattice: a two dimensional regular lattice of rectangular sites, where each site may be empty with probability p or occupied with probability q=1-p. The attractive feature of this model is the imprinting of the city structure in the city probabilistic description. This presentation highlights merits and limitations of the two approaches. For the DGM a numerical code is discussed, and examples of computations are critically shown, as well as possible extensions of the software. For the SEM analytical results are presented, leading to interesting theoretical considerations. In both cases, some experimental results are referred, too. Alternative models worth of future exploration are suggested.

14:00-17:30 Students Presentations
Matteo Lanati, University of Pavia, Italy, “End to End QoS in a wired and wireless network”
Cosimo Palazzo, University of Lecce, Italy, “An IPSec-aware Protocol for TCP Freezing in Satellite Networks”
Maguolo Federico, University of Padova, Italy, “VoIP over WLANs: performance evaluation and improvements”
Marco Levorato, University of Padova, Italy, “Cross-Layer Design of MIMO Ad Hoc wireless Networks with Spatial Multiplexing and Decision Feedback Multiuser Detection”

Tuesday June 27, 2006

9:00-12:30, Tutorial Session
Prof. Gabriella Olmo, Polythechnic University of Torino, Italy: ‘Multiple description coding of visual data
Abstract: Multiple description coding is recognized to be an effective tool for resilient transmission of multimedia data over networks subject to packet erasures. The basic principle is to split the data into multiple streams, or descriptions, which can be independently decoded and yield independent contribution to the quality of the recovered signal. In this tutorial, the most popular MDC techniques for image and video data are described, and the related pros and cons are discussed and compared with single description and layered coding. Particular attention is devoted to MDC compatible with the JPEG 2000 and H.264/AVC standard co-decoders.

14:00-17:30 Students Presentations
Ottavio Campana, University of Padova, Italy, ”Multiple Description Video Compression with MDSQ and H.264/AVC”
Daniele Menon, University of Padova, Italy, “Techniques for demosaicing in digital cameras”
Francesco Pecile, University of Udine, Italy, “Synchronization Algorithms for Multiuser Filtered (FMT) Systems”
Matteo Trivellato, University of Padova, Italy, “Joint Timing/Frequency Synchronization and Pilot aided Channel Estimation in Time-Varying Multipath Fading Channels for IEEE 802.11p Transmissions”

Wednesday June 28, 2006

9:00-12:30, Tutorial Session
Prof. Marco Lops, University of Cassino, Italy: ‘Finite Random Set Theory: a Natural Framework for Multi-User Communications
Abstract In a typical multiple-access communications scenario, the number of active users, their location, as well as the parameters that characterize their channels vary with time . Thus techniques aimed at identifying not only the transmitted data, but also the users parameter play a central role in analysis and design of wireless transmission systems. Examples of applications e.g., in Multi-User Detection, spatial multiplex schemes and ad hoc networks. 1)MUD receivers must account for the number of interferers being active at any given time. Adaptive systems, wherein the channel is first sensed, and the receiver then adjusted to the real scenario have been proposed, e.g., in [1], while more and more sophisticated procedures have been analyzed in [2-3], in order to deal with more and more challenging scenario. 2)In spatial multiplexing schemes, the total system throughput may be optimized by identifying a subset of users to which the power is allocated (see e.g. [4]). 3)In ad hoc networks, optimal transmission strategies require the identification of active nodes in the neighbourhood of the transmitter.
In this talk, considering, as a field of application, adaptive MUD, the issue of joint identification and parameter estimation of active users is re-formulated in a completely different context, the so-called Finite Random Sets Theory (FRST). To fix the ideas, assume that a classical CDMA system, and denote by N the maximum user number. Assuming a constant load in the network (e.g., K <= N active users) is definitely unrealistic, since it may happen that some of the users become silent at a the end of a transmission phase, while some other users start transmitting. Since each user is characterized by a flag, containing its identity (an integer number <= N), an information symbol, plus a number of unknown parameters (e.g., average power, phase, and so on), the set Xt of users active at time “t” is random, not only in its cardinality, but also in its elements, each element being a members of a hybrid space, S say, the Cartesian product between a countable set U and the space R^d. Additionally, denoting by S_t the random set of the users surviving into epoch t+1 and by N_t the set of newly born users, it is understood that: X_{t+1}=S_t U N_t. Just to give an idea, being able to estimate (possibly in a recursive way) the set sequence {X_t} would enable undertaking channel sensing and user demodulation in a unique step, and would offer the possibility of accounting for a number of effects, such as the users’ activity factor, the inherent death-and-birth process of the users themselves, and so on. To this end, several intermediate steps should be taken, i.e.: Develop a suitable probability space so that “set probability densities” can be defined; Devise manageable models for the evolution of the random set Xt under diverse instances of channel state information; Derive and implement structures which optimally solve the detection/estimation problem above.
The course develops, obviously in a tutorial form, the three points above, and is mainly intended to introducing FRST as a tool whose applicability appears to reach well beyond the issue of adaptive multi-user detection. A tentative schedule might be as follows: Review of classical adaptive multi-user detection [1-3] and reformulation in terms of random sets, with a brief history of FRST: 1 hour; Finite Random Set Statistics: belief functions, set densities, set derivatives, set integrals, set estimation [5]: 1 hour; Multi-User Detection in a Static Channel; Multi-User Detection in a dynamic Channel: set sequence detectors and Bayes Recursions [6-7]: 1 hour; Conclusions and hints for further developments: 0.5 hours.
[1] U. Mitra, V. H. Poor, “Activity Detection in a multi-user environment”, Wireless Personal Communications, vol. 3, No. 1-2, pp. 149-174, January 1996.
[2] K. W. Halford, M. Brandt-Pearce, “New user Identification in a CDMA system”, IEEE Transactions on Communications, Vol. 46, No. 1, pp. 144-155, January 1998.
[3] T. Oskiper, V. H. Poor, “Online Activity Detection in a Multi-User environment using the matrix CUSUM algorithm”, IEEE Transactions on Information Theory, Vol. 48, No. 2, pp. 477-493, February 2002.
[4] W. Yu, Y. Rhee, “Degrees of freedom in multi-user spatial multiplex with multiple antennas”, submitted for publication, 2004.
[5] I. R. Goodman, R. P. S. Mahler, H. T. Nguyen, Mathematics of Data Fusion, Dordrecht, The Netherlands: Kluwer, 1997.

14:00-17:30 Students Presentations
Livio Denegri, University of Genova, Italy, “Propagation Model in Urban Street Canyon Environment”
Marco Pini, Polythechnic University of Torino, Italy, “Innovative Signal Processing Techniques for the Quality Monitoring of Global Navigation Satellite Systems (GNSS) Signals”
Daniele Veronesi, University of Padova, Italy, “Multiple Frequency Offset Compensation in Cooperative Wireless Systems”
Federico Boccardi, University of Padova, Italy, “The p-Sphere Encoder: Peak-Power Reduction by Lattice Precoding for the MIMO Gaussian Broadcast Channel”

Thursday June 29, 2006

9:00-12:30, Tutorial Session
Prof. Giorgio Baccarani, University of Bologna, Italy, ‘Trends and Challenges in VLSI Technology Scaling
Abstract: This tutorial addresses the evolution of micro- and nano-electronics technology towards increasing integration levels. First, the functional requirements of CMOS logic gates are briefly summarized, and the main tradeoffs for performance optimization are discussed within the constraints posed by noise immunity and power dissipation. Next, the generalized scaling laws are derived and the fundamental scaling limits of semiconductor devices are discussed, with special emphasis on static and dynamic power dissipation. New device architectures, such as double-gate (DG) and nano-wire (NW) FETs are then illustrated, to alleviate the short-channel effect and improve the overall device performance. Next, the growing problems of interconnect-related gate delay and noise immunity are addressed along with system-level aspects. Finally, device modeling issues are addressed, with special attention devoted to the coupled Schrödinger-Poisson equations and to the required transport models, such as the quantum drift-diffusion (QDD) and quantum-ballistic (QB) models.

14:00-17:30 Students Presentations
Franco Zanon, University of Padova, Italy, “Detrimental Effects Induced by Alpha Irradiation at 2 MeV on AIGaN/GaN High Electron Mobility Transistors”
Lorenzo Trevisanello, University of Padova, Italy, “Degradation mechanisms of GaN/InGaN light emitting diodes during thermal aging”
Nicola Montemezzo, University of Padova, Italy, “Brokaw and Kuijk Bandgaps Susceptibility to RF Interferences: Measurements, Modeling and Provisions”
Andrea Maniero, University of Padova, Italy, “A 0.35 µm SiGe Low-Noise Amplifier for UWB Receivers with Integrated Interferer Rejection”

Friday June 30, 2006

9:00-12:30, Tutorial Session
Prof. Giuseppe Bianchi, University of Rome ‘Tor Vergata’, Italy, ‘Performance analysis of 802.11 wireless LAN networks
Abstract: Performance analysis of 802.11 (CSMA/CA) wireless LAN networks has envisioned a renewed and extensive research interest in the last few years. Traditional modelling methodologies, developed in the 70s and 80s were bound to special assumptions, such as, e.g. slotted operation, and/or colliding traffic modeled as a Poisson process, which somehow limited their practical application. Conversely, modern analytical models have been shown to be able to accurately account for important implementation details such as the specific Binary Exponential Backoff rules employed in the practical implementation of CSMA/CA (namely, the Distributed Coordination Function – DCF – of the IEEE 802.11 standard). Goal of this talk is to review the basic foundations and principles of these new modeling methodologies, and to show how they have been successfully applied not only to model single-cell networks, but also how they have been generalized to capture and thoroughly explain critical issues such as lack of fairness and flow starvation emerging in multi-hop/Mesh environments. Rather than providing a report on this research area, the talk wishes to provide an understanding of the basic key ideas, principles, and methodologies which, we believe, are not bounded to the specific framework of Wireless LAN systems, but may be successfully applied to other systems and protocols.
Detailed outline of the talk: – Introduction, review of 802.11 DCF, basic modeling principles – Throughput bounds for 802.11 networks – Markov-Chain saturation throughput analysis and its reinterpretation/generalization in terms of elementary probability theory – Model extensions for error-prone channels and non saturated conditions – Models for multi-hop networks

14:00-17:30 Students Presentations
Chiara Buratti, University of Bologna, Italy, “On the Number of Cluster Heads Minimizing the Error Rate for a Wireless Sensor Network Using a Hierarchical Topology over IEEE802.15.4”
Virginia Corvino, University of Bologna, Italy, “Scheduling of Mixed Traffic over MC-CDMA under Varying Load and Channel Conditions”
Simone Merlin, University of Padova, Italy, “Resource Allocation in a 4th Generation Wireless Cellular System”