BIG DATA SENSING COMPRESSION and COMMUNICATION

ING-INF/03 - 9 CFU - 79 hours

Coherence of the course with reference to the Master Degree in “Telecommunications Engineering”

Data is growing and has grown very fast in the last years.”Big Data” analytics is challenging today because of the unprecedented large data volumes. In this course, we will introduce at first the digital transmission theory which applies independently of the specific type of data; then we will detail the structure of data generated in big data sensing applications, by distinguishing the type and structure of data. In the following we will discuss SoA methodologies which can be used to compress this data based on its intrinsic features; finally, communication protocols for remotely delivering this data will be described and detailed. In this way students will be provided with communication engineering competences allowing them to actively communicate with experts in various fields by providing focused and competent data analysis for every application, such as in scientific, technological or business fields. Students will also be able to exploit the competences gained for design processes of collection, compression and communication of heterogeneous big data.

Learning Objectives

The course aims to provide students with some basics of digital transmisison, information generation, encoding, compression and communication for big data scenarios.

Dublin Descriptors

1. Knowledge and understanding (Conoscenza e capacità di comprensione) - The course aims to provide students with knowledge and understanding of techniques and algorithms for acquisition and processing of data (e.g. sensor generated data, images, audio files) collected in smart environments such as in environmental monitoring, e-health, smart cities and/or vehicular scenarios. Then students will understand and study techniques for data compression both at the sources and, in a distributed way, in the network. Finally technologies and architectures for the transmission of big data will be studied.
2. Applying knowledge and understanding (Capacità di applicare conoscenza e comprensione) - After attending this course, students will be able to manipulate, process and reconstruct different types of data acquired from a smart environment, design compression algorithms suitable to perform data compression both at the data sources or into the network, choose and exploit the most appropriate set of technologies for data transmission in big data scenarios. Finally students will be able to solve specific big data design problems in realistic scenarios.
3. Making judgements (Autonomia di giudizio) - Upon completion of the course the students will gain independent and critical understanding skills as well as ability to discuss design aspects in real big data scenarios, commenting also on the design choices. Finally, at the end of the course, the students will be able to prosecute independently their study of other engineering-related disciplines with the ability to appropriately use big data design considerations in the appropriate context.
4. Communication skills (Abilità comunicative) - Students attending this course will learn to communicate and discuss/describe relevant Big Data application scenarios. Also they will be able to critically discuss and illustrate the most relevant design aspects to be taken into account upon focusing on generation, elaboration and communication of huge amounts of heterogeneous data like those generated in IoT networks.

The course consists of lectures and laboratory activity. The theorethical lectures are taught by the teacher while laboratory activities, consisting of exercises, will be carried out in collaboration by the teacher and by the students who are invited to solve, with the support of the teacher, exemplary problems. In addition, other lectures will be devoted to the illustration of software tools, e.g. Mathworks Matlab, useful for the solution of specific problems.

In case the course is taught in mixed or only remote mode, it could be needed to implement some necessary variations as compared to what has been foreseen and reported in the syllabus.

Prerequisites: Basics of maths (integrals, derivatives, matrixes, vectors, functions, scientific/exponential notation), basics of communication systems (not strictly required).

Attending classes is not mandatory but strongly recommended.

The final exam will consist of a colloquium with the teacher on the topics dealt during the course. The final exam could be eventually done remotely in case of COVID contingency.

Introduction (approx 3 hours): Introduction to Internet of Things-Introduction to big data-Definition of big data-Types of big data-operations on big data-Examples of big data.

Part 1 (approx 40 hours): Digital transmission: Space of signals, Vector representation of signals, Gram-Schmidt orthogonalization -Survey on main analog and digital modulation techniques, Error probabilities for different modulation schemes, MAP criteria for signal reconstruction, Transmission on Gaussian channel -Viterbi decoding, convolutional and Trellis codes -Minimum bandwidth systems, Maximum likelihood estimation, receiving systems -Fading channels, Clarke hypothesis, multipath, Rayleigh and Rice fading -Equalizers, Zero Forcing equalizers,

Part 2 (approx 12 hours). Big data sensing: Types of data - Audio sources - Basics of acoustics - Human earing fundamentals - Basics of digital audio - Digital encoding - Sampling Theory - Different audio file formats - Compressed audio - Video sources - Basics of video encoding - Different video file formats - Multimedia transmission - Fundamentals - Jitter and synchronization - Multimedia file formats - Data sources - Data file formats - Examples of different mechanisms for data generation.

Part 3 (approx 10 hours). Big data compression: Source coding - Compressive sensing - Channel coding - Examples of compression techniques applied to different types of data.

Part 4 (approx 15 hours). Big data communication: Technologies for the IoT - WiFi - LoRa - SigFox - Examples of communication between nodes exploiting some of the technologies discussed above.

The following texts are suggested readings. During the course, the teacher can also suggest further readings (e.g. scientific papers and articles) on specific topics.

-A. Rezzani. Big Data Analytics: Il manuale del data scientist, Apogeo Maggioli Editore

-V. Lombardo, A. Valle. Audio e multimedia, 4th edition, Apogeo Maggioli Editore.

-Z. Han, H. Li, W. Yin. Compressive sensing for wireless networks. Cambridge University Press.

-F. Wu. Advances in visual data compression and communication: Meeting the Requirements of New Applications, CRC Press.

-U. Mengali, M. Morelli, Trasmissione numerica, Mc Graw Hill