Budapest University of Technology and Economics, Faculty of Electrical Engineering and Informatics

    Belépés
    címtáras azonosítással

    vissza a tantárgylistához   nyomtatható verzió    

    Perception and Signal Processing

    A tantárgy neve magyarul / Name of the subject in Hungarian: Érzékelők és jelfeldolgozás

    Last updated: 2023. október 5.

    Budapest University of Technology and Economics
    Faculty of Electrical Engineering and Informatics
    MSc Electrical Engineer
    Course ID Semester Assessment Credit Tantárgyfélév
    VIMIMA20   2/1/0/v 5  
    3. Course coordinator and department Dr. Orosz György,
    4. Instructors Dr. György Orosz, associate professor, BME MIT
    7. Objectives, learning outcomes and obtained knowledge This course gives insight into the perception of physical signals and methods of pre-processing of signals in embedded systems. The course introduces the most commonly used sensors and the disturbing and distorting effects of observing the environment. Several basic signal processing methods will be introduced which can be used in lot of application fields. These sensors and signal processing methods form the basics of modern intelligent systems.
    8. Synopsis

    LECTURES:

    1. Introduction to embedded systems: sensing the environment, processing data. Investigating a typical signal path: sensors, perception and signal conditioning, sampling, data processing. Typical embedded signal processing architectures: microcontroller, DSP, FPGA, GPU.

    2. Temperature sensors: e.g., thermocouple, NTC/PTC, infra, semiconductor. Measurement of light intensity, e.g. photodiode, phototransistor, photo resistor, photovoltaic. Specific properties and applications of sensors.

    3. Vibroacoustic sensors: MEMS and conventional analog sensors (piezo, electret, geophone). Charge and voltage output devices, signal conditioning issues, typical specifications.

    4. Measurement of position, displacement, rotation: incremental transducers, LVDT, optical sensors, time-of-flight sensors, Hall-sensors and magneto-resistive sensors, inductive sensors. Force and torque measurement: strain gauges, piezo, force-sensitive resistor.

    5. Measurement of current: shunt resistance (bottom and top), current transformer/Rogowski coil, magnetic field based (Hall sensor, fluxgate, magneto-resistive) sensors. Measurement of ECG and photoplethysmographic signals.

    6. Interpretation of DFT for periodic and stochastic signals. Calculation of equivalent noise bandwidth, signal to noise ratio. Coherent/non-coherent sampling, distortion effects. Alternative interpretations of the discrete Fourier transform (DFT): matrix transform, filter bank, LS estimation (generalization to sine components of arbitrary frequency).

    7. DFT applications: convolution acceleration, real DFT computation using complex DFT, cepstrum computation. Wavelet transform, description of wavelets, implementation. Discrete cosine transform.

    8. Classification of digital filters. Overview of the properties of IIR and FIR filters. Amplitude and phase characteristics. Types of filters: FIR: LS and smooth wave; IIR: Butterworth, Chebyshev, elliptic, Bessel-Thomson.

    9. Filter design procedures for FIR and IIR filters (LS, Parks-McClellan, windowing, bilinear transform, pulse invariant transform). Realization forms of digital filters, biquad implementation.

    10. Specialties of fixed point fractional representation, performing operations with fixed point fractional numbers, design difficulties. Nonlinear filters and outlier detection: median filter and variants, Hampel filter.

    11. An overview of problems stemming from different sampling frequencies. Implementation of decimation and interpolation in time and frequency domains. Decimation and interpolation filter design, polyphase filter. Polynomial interpolation.

    12. Numerical optimization problems: root locus search, extreme value search. Formulating a mathematical problem, types of cost function, interpretation. One and multi-parameter problems, conditional search for extreme values. Methods using first and second order derivatives.

    13. Optimization methods based on different heuristics. The problem of local extrema. Convergence problems, ill-conditioned cases. Illustration of numerical optimization using the Least-Mean Square (LMS) algorithm as a data-invariant filter for real-time embedded systems.

    14. Consultation, extra space for missing lessons due to holidays, examples of applications.

     

    PRACTICES:

    1. Characterization of systematic and random static errors: error analysis of complex circuits, contribution of different passive and active components to the error and non-linearity of the signal path. Reducing the errors with special components and designs. 

    2. Noise analysis: noise of circuit components (amplifiers, passive components, power supplies), noise suppression/immunity, ambient noise/disturbance, shielding, design solutions. Jitter and its effects.

    3. Introduction to some basic signal processing steps through a real application. Block and sample processing: difference between real-time embedded processing and offline processing. Interpretation of timings. Implementing simple signal processing algorithms, e.g. measuring signal parameters (frequency, amplitude...).

    4. Introduction of filter design software packages, design and test FIR and IIR filters with different specifications on real signals.

    5. Spectrum analysis examples. Fault detection in embedded systems based on spectral imagery: independent periodic, modulating errors, random errors/jitter, distortion, and their derivation from software and hardware properties of embedded systems.

    6. Examples for designing decimating and interpolating filters for different tasks. Synchronization of different sampling frequencies in distributed embedded systems.

    9. Method of instruction 2 hours of lectures per week, 1 hour of practice (computational exercise and computer laboratory exercise). 
    10. Assessment

    During teaching period: midterm exam.

    In the exam period: written exam.

    11. Recaps Mid-term exam can be once repeated during teaching period.
    12. Consultations On request, upon agreement with the lecturer in advance.
    14. Required learning hours and assignment
    Contact lesson 42
    Preparation for lessons  13
    Preparation for mid-term exam  25
    Homework  0
    Independent studying  20
    Preparation for exam  50
    Total 150
    15. Syllabus prepared by Dr. György Orosz, associate professor, MIT