B8378 - SENSORI ELETTRONICI INTELLIGENTI LM

Learning outcomes

The course is designed to provide the fundamentals of sensing system design from a unified perspective, going beyond the idea of sensors as merely a collection of technologies. Sensor design will be introduced as the implementation of a predictive model across the entire information chain, from the electronic interface to the machine learning classifier. Students will learn how to approach sensor design by identifying key constraints and trade-offs, and by selecting the most effective implementation technologies. The course will also feature practical examples of embedded machine learning sensors. Lectures are often based on a bottom-up approach to better understand principles from examples.

Course contents

PART I (6 CFU)

Program for:

• SENSORI INTELLIGENTI [Laurea Magistrale in Ingegneria Elettronica e dell'Informazione]
• SENSORS & NANOTECHNOLOGY [Laurea Magistrale in Biomedical Engineering]
• EMBEDDED SENSORS [Laurea Magistrale in Ingegneria e Scienze Informatiche]

The goal of Part I is to give students fundamental know-how in sensor system design, approached from the perspective of behavioral building blocks and their technological implementation.

FUNDAMENTALS

· Introduction of the course, example of research.

· Fundamentals on sensors. Signals and information. Sensor as an information classifier. Interface, raw data and signal processing (Machine Learning).

· Basics on measurements. Characterization and operating mode. The concept of distributions and estimation. DC measurements. Digits and full-scale. Random and systematic errors. Introduction to the resolution, precision and accuracy concepts. The role of time and AC measurements. Power spectral distributions.

· Ideal sensor modeling. Interface as a black-box. The quasi-static function and its limits. The concept of sensitivity. Time and space quantization, the A/D converter. Acquisition chains and gain plots. Signal bandwidth.

· Recall of mathematical tools. Random signals, power and correlations. Random and ergodic processes.

· Non-ideal sensor modeling. The resolution/uncertainty duality. The law of propagation of errors.

· Comparing signal with noise. The concept of minimum detectable signal. The signal-to-noise ratio and the dynamic range. Systematic error modeling.

· The quantization process. The concepts of number of resolution levels and effective number of bits. Relationship with Information Theory. Resolution degradation in acquisition chains. Example of designs. The role of bandwidth, oversampling and non-aliasing filters.

· Feedback sensor design with examples.

· Sensor trade-offs and example of design.

ELECTRONIC INTERFACE DESIGN

[This section is OPTIONAL and NOT REQUIRED for EMBEDDED SENSORS students]

· Introduction of noise in circuit design. PSD and noise density. Thermal noise shot noise ad flicker noise. KTC noise. Superimposition of noise powers in uncorrelated signals. Equivalent noise bandwidth. Concept of input-referred noise. Input referred noise in BJT and MOS devices. Noise in OPAMPs.

Resistance sensors with examples. Strain-gauges. RTDs e PRTs.Thermistors, NTC e PTC. Magnetic sensors.

· Capacitive sensors. Kelvin guard ring. Charge amplifier. Differential capacitive sensing. Capacitive accelerometers. Noise in charge amplifiers. Correlated double sampling (CDS).

· Lock-in e chopper sensing. Complex impedance measurements by lock-in sensing.

· Introduction to optical sensors. The photodiode. Charge and voltage photodiode readout in storage mode. Passive pixel CMOS sensors (PPS) and active pixel (APS) sensors. APS with correlated doble sampling.

· Sensor networks.

SENSORS ECOSYSTEMS

· Brief overview of embedded sensor networks. I2C and SPI.

· Brief overview of wireless sensor communications and networks.

TOWARDS MACHINE LEARNING EMBEDDED SENSORS

· The concept of predictive models. Measured vs. predicted measurements plots.

· Dealing with multiple variables: scatter plots and correlation matrix in multivariate analysis.

· Reduction of variables using decomposition techniques. The PCA.

· Reduction to output values, introduction to the PLS technique

· Example of application: embedded spectral sensors

PART II (NANOTECHNOLOGY SECTION, 3 CFU)

Program for:

• SENSORS & NANOTECHNOLOGY [Laurea Magistrale in Biomedical Engineering]

Part II is offered exclusively to students in Biomedical Engineering and is intended to introduce the physical foundations of biosensor transduction principles at the micro- and nanoscale. These concepts will then be applied in the context of protocols and instruments for biomedical laboratories. This part will also include a hands-on laboratory experience at the end of the course.

· Photons, electrons and energetic interactions with matter. Fluorescence, phosphorescence and Raman scattering. Photon-electron transduction. Black body emission. Quantum efficiency and spectral sensitivity of transduction in semiconductors. Responsivity.

· Basics of geometrical optics. Refraction and diffraction. Optical resolution and its limits. Optical microscope. Basics of electron microscopy. Basics of atomic force microscope (AFM).

· The origin of noise. Brownian noise. Thermal noise and its derivation. Poisson processes. Shot noise and related derivation. The flicker noise and its derivation. Physical origin of flicker noise.

· Introduction of biosensing principles. Ion-electron transduction. Electrical polarization of the matter. Metal-liquid interface and interfacial states. Electrical models of ion/electron interfaces and electro impedance spectroscopy.

· Introduction to nanotechnology. Nanosensors: nanopores, nanotubes, nanowires, graphene, nanodots.

· Lab experience on electro-impedance or Raman spectrometry.

Readings/Bibliography

Main texbook

M. Tartagni, Electronic Sensor Design Principles, Cambridge Press, 2021

------------------------------------------------------------------------------------

Physical principles:

R. Feynman et al., The Feynman Lectures on Physics, Addison Wesley, 1963

Noise in electronics:

P. Gray, R. Meyer, Analysis and Design of Analog Integrated Circuits, Wiley 1993

Ion/electron interface, micro- and nano-fabrication:

H. Morgan, N. Green, AC Electrokinetics: colloids and nanoparticles, RSP Press, 2001

M. Madou, Fundamentals of Microfabrication, CRC Press, 2002

Electronic interfaces:

D. Johns, K. Martin, Analog Integrated Circuit Design, Wiley, 1997

B. Razavi, Design of Analog CMOS Integrated Circuits, McGraw-Hill, 2000

Machine learning and multivariate analysis:

Kevin Dunn, “Process improvement using data”, [Online]. Available: https://learnche.org/pid/

K. P. Murphy, Probabilistic machine learning: an introduction. Cambridge, Massachusetts: The MIT Press, 2022.

Teaching methods

The course is based on face-to-face lectures given by the instructor, supported by slides that will follow the syllabus and be available on Virtuale. The Biomedical Engineering part provides a brief laboratory experience.

Assessment methods

The exam consists of an oral interview lasting 40–60 minutes covering both teaching modules. Particular emphasis will be placed on the student’s critical abilities regarding the topics of the syllabus, as well as on their capacity to contextualize and articulate concepts. The exam may be taken in either Italian or English. In both cases, the appropriateness of the terminology used by the student will be assessed.

 

 

Students with Specific Learning Disorders (SLD) or temporary/permanent disabilities are advised to contact the University office in charge in advance, at the address (https://site.unibo.it/studenti-con-disabilita-e-dsa/it ). The office will be responsible for proposing any adaptations, which must in any case be submitted, at least 15 days in advance, for the approval of the instructor, who will assess their suitability in relation to the learning objectives of the course.

Teaching tools

Slides will be available on Virtuale.

Office hours

See the website of Marco Tartagni