Lorenzo Favalli

Object tracking for retrieval applications in MPEG-2

The work presented in this article describes a tool for object tracking, notes insertion, and information retrieval, applicable to MPEG-2 sequences. Maximum compliance with the MPEG standard is sought, so the added information is transmitted as side information without affecting the actual video-audio stream as defined in the MPEG-2 standard. Additional processing is added to a standard sequence, allowing for automatic tracking of one object across different groups of pictures. Results show that the proposed algorithm is capable of tracking objects with a good degree of precision. Features are included to alert the human operator when objects disappear, or must be considered lost, due to an excessive change in their shape.

Multipath channel modeling with chaotic attractors

Previous works have introduced models based on deterministic chaos equations, aiming to simulate transmission processes over mobile radio digital channels at the level of the post detection error stochastic process and to reproduce their renewal or nonrenewal behaviors. In particular bursts and clusters of errors are generated by these models, exploiting the correlation between successive points sampled from the trajectories of suitable strange attractors. In this paper chaos equation models are reviewed and relevant results are discussed. Different approaches have been tested, using one or more attractors, and calling for light or heavy preprocess procedures to cope with the statistical characteristics of target gap-time series. Models derived from hidden Markov chains have been implemented to provide comparison of results. Target sequences were supplied by simulation of digital enhanced cordless telecommunications (DECT) channels or derived from reception of DECT signals by a mobile unit in indoor environments.

A suboptimal approach to channel equalization based on the nearest neighbor rule

Applications of clustering and neural network techniques to channel equalization have revealed the classification nature of this problem. This paper illustrates an implementation of a global system for mobile communications (GSM) receiver in which channel equalization and demodulation are realized by means of the nearest neighbor (NN) classifier algorithm. The most important advantage in using such techniques is the significant reduction in terms of the computational complexity compared with the maximum likelihood sequence estimation (MLSE) equalizer. The proposed approach involves symbol-by-symbol interpretation and the knowledge of the channel is embedded in the mapping process of the received symbols over the symbols of the training sequence. This means that no explicit channel estimation need be carried out, either with correlative blocks or using neural networks thus speeding up the entire process. The performance of the proposed receiver, evaluated through a channel simulator for mobile radio communications, is compared with the results obtained by means of a 16-state Viterbi algorithm and other suboptimal receivers. It is shown that the presented algorithm increases the bit error rate (BER) compared with the MLSE demodulator, but the performance degradation, despite the simplicity of the receiver, is kept within the limits imposed by the GSM specifications.

A coding and retransmission protocol for mobile radio data transmission

Mobile radio systems suffer from severe channel impairments. As services are moving toward a full extension of those available over fixed networks, new coding strategies are being developed to support data traffic beside the more usual voice communication. This paper presents a modified type II hybrid FEC/ARQ protocol in which each packet to be transmitted is decomposed into sub-packets each of which can be independently tested for correctness and retransmitted if necessary. The complete coding and transmission scheme is designed to be compatible with the frame format used in GSM networks. The performances of the proposed algorithm are evaluated by computer simulation using the system specifications contained in the ETSI-GSM recommendations. Throughput comparison with a traditional selective repeat (SR) protocol is presented to support the robustness and effectiveness of the presented scheme.

Modeling and analysis of aggregate and single stream Internet traffic

We analyze traffic traces obtained from measurements at the edge gateway of the campus network of the University of Pavia. These traces have been obtained simply using the tcpdump utility and then analyzed and modeled using two different approaches. The first is based: on a hidden Markov model (HMM) the other on a stochastic generator based on a chaotic attractor. It is shown that the HMM is capable of capturing the general statistics of the observed streams despite poor long term correlation characteristics, while the chaotic model can achieve somewhat better results in long term characteristics and somewhat worse in short term. Moreover, the chaotic model is sensitive to the parameter extraction procedure.

Cognitive radio resource management exploiting heterogeneous primary users and a radio environment map database

The efficient utilization of radio resources is a fundamental issue in cognitive radio (CR) networks. Thus, a novel cognitive radio resource management (RRM) is proposed to improve the spectrum utilization efficiency. An optimization framework for RRM is developed that makes the following contributions: (i) considering heterogeneous primary users (PUs) with multiple features stored in a radio environment map database, (ii) allowing variable CR demands, (iii) assuring interference protection towards PUs. After showing that the optimal solution is computationally infeasible, a suboptimal solution is consequently proposed. Performance evaluation is conducted in terms of total achieved data rate and satisfaction of CR requirements.

Cognitive Radio Resource Management Exploiting Heterogeneous Primary Users

In this paper, a novel Cognitive Radio Resource Management (RRM) is proposed to improve the spectrum utilization efficiency. In this system, heterogeneous Primary Users (PUs) with multiple features are considered where these PU features are exploited to improve the adaptability in Cognitive Radio (CR) networks and, thus, to design an efficient Cognitive RRM. An optimization framework is developed by considering heterogeneous PUs and variable CR demands while assuring interference protection towards PUs. A suboptimal solution is proposed after showing that an optimal solution is computationally infeasible. Simulation results are conducted in terms of total achieved data rate and satisfaction of CRs requirements.

Satellite channels modelled by chaotic bit error generators

A comparison is carried out between bit error generators based on chaotic models and markovian ones. Given an Error Free Run characteristic to shape the probability distributions of the gap length generated by a strange attractor, renewal and nonrenewal behaviors of the error time series and other statistical parameters can be obtained with some degree of freedom. Some parameter matching is possible with time series generators based on markovian chain.

Recovery of ISI channels using multiresolution wavelet equalization

Equalization of ISI is usually performed in the time domain. In fact it is difficult to operate in the frequency domain, since with the usual spectral techniques based on the Fourier transform, correspondence between temporal and frequency components is lost in the transformation. In this paper we present an application of the wavelet transform that exploits the joint time-frequency representation of this transform to equalize channels affected by ISI. It is shown that using the description of wavelets as filters, and exploiting the multispectral decomposition of the received signal, ISI can be dramatically reduced.

ILPS: a scalable multiple description coding scheme for H.264

The most recent literature indicates multiple description coding (MDC) as a promising coding approach to handle the problem of video transmission over unreliable networks with different quality and bandwidth constraints. In this work, we describe an approach that moves from the concept of spatial MDC and improves coding efficiency by exploiting some form of scalability. In the algorithm, we first generate four subsequences by sub-sampling, these subsequences are then taken in pairs that will form each of the two descriptions. For each description, one of the original subsequences is predicted from the other one via some scalable algorithms, focusing on the inter layer prediction scheme. The proposed algorithm has been implemented as pre- and post- processing of the standard H.264/SVC coder. The experimental results presented show that the algorithm provides excellent results in several conditions including 3D video sequences.