Signal processing
[ELEC2900]
[30h + 30h] 5 credits
This course is taught in the 2d semester.
Teacher(s): Benoît Macq, Luc Vandendorpe
Language: English
Level: Second
Aims
- make the link between the analog description of sampling and sequences,
- modify the sampling rate of a discrete time signal i.e., upsample or downsample lowpass or passband signals, deterministic or random; implement these operations by means of efficient structures, in particular polyphase structures,
- understand the consequences of sampling the spectrum,
- design from a spectral template, finite impulse response (FIR) filters by means of different optimum and suboptimum methods,
- design from a spectral template, infinite impulse response (IIR) filters; understand and use the bilinear transform; design filters based on criteria discussed in "INMA2731 : Processus stochastiques",
- design systems for processing multidimentional signals, in particular images,
- understand and use linear transformations for decorrelation, multiresolution analysis, and sicriminant analysis
Main themes
Content and teaching methods
- Sampling rate conversion : interpolation, downsampling, lowpass and bandpass signals, deterministic and random signals,
- Structures and graph theory (introduction), polyphase components,
- Discrete Fourier transform,
- Finite impulse response filters,
- Basics of analog filters and templates,
- Bilinear transform and design of infinite impulse response filters
- Processing of random signals,
- Processing of multidimensional signals,
- Denoising and singularity detection,
- Orthogonal transforms,
- Decorrelative transforms,
- Wavelet transform,
- Linear discriminant transform,
- Non parametric (periodogram) and parametric (process identification) spectral analysis
Other information (prerequisite, evaluation (assessment methods), course materials recommended readings, ...)
There will be lectures interleaved with practical training (in teaching room or computation center with MATLAB)
Prerequisites :
INMA1731 : Random processes : estimation and prediction
Assessment :
Written examination about exercices, with notes
Could be given in English