Overview
Module 4 introduces digital signal processing that can push the system to more closely resemble an additive white Gaussian noise (AWGN) channel that has been analyzed in the previous modules. PartA of module 4 examines equalization techniques to be applied at the receiver. For linear equalization, the matched filter is replaced with a filter that reduces the distortion as opposed to simply maximizing the signal to noise ratio. Nonlinear techniques are also addressed. Equalizers include the minimum mean square error (MMSE), zero forcing (ZF), maximum likelihood sequence estimator (MLSE) and decision feedback equalizer (DFE).
Videos
Course notes
Exercises
Objectives
- appreciate the role of an equalizer
- know the taxonomy of equalizers
- know the characteristics of the channel that influences the complexity, design and performance of an equalizer
- be able to discuss the characteristics, advantages and disadvantages of different equalizers
- know several sources of ISI in a communications system
- know the operating principle a maximum likelihood sequence estimator (MLSE)
- know the stengths and weaknesses of MLSE
- know the architecture of linear transversal filters
- know definition of a ZF equalizer, both in the frequency and temporal domains
- know the information required for finding the ZF equalizer, both in the frequency and temporal domains
- know the definition of an MMSE equalizer
- be able to contrast the ZF and MMSE equalizers, in complexity, information required to exploit it, optimality, and weaknesses
- know strategies to make equalizers (ZF or MMSE) adaptive via training sequences and the costs and benefits of the adaptive version
- contrast the two types of adaptation: "data-aided" and "decision-directed"
- know the block diagram and operating principle of a DFE (decision feedback equalizer)
- contrast the two types of decision-making strategies: "decision-directed" and "decision feedback DFE"
Reading
- section 3.4 Equalization
- Sklar's article: How I Learned to Love the Trellis
- Ziemer and Tranter, Principles of Communications: extracts of section 5.5 Zero-Forcing Equalization and section 9.9 Equalization
Overview
Module 4 introduces digital signal processing that can push the system to more closely resemble an additive white Gaussian noise (AWGN) channel that has been analyzed in the previous modules. PartB of Module 4 examines signaling strategies at the transmitter that can avoid distortion from a band-limited channel. These strategies reduce the signal processing burden for equalization. They are designed to avoid intersymbol interference (ISI). Methods include orthogonal frequency division multiplexing (OFDM) and discrete multitone (DMT).
Videos
Course notes
Exercises
Objectives
- know the link between multipath channels and frequency-selective fading
- be able to explain how ISI (and equalization) is avoided with OFDM
- be able to explain OFDM's robustness against fading
- know block diagram for OFDM transceivers
- be able to explain the role of the inverse Fourier transform for OFDM transmission
- know OFDM's peak-to-average problem
- know the advantages and disadvantages of OFDM vs. equalization
- contrast OFDM and DMT approaches; how they are similar and different
- be able to explain the optimization made for DMT
Reading
- Bahai, Saltzberg and Ergen, Multi-carrier Digital Communications: extracts of section 3.3 Bit Allocation and 3.4 Bit and Power Allocation Algorithms for Fixed Bit Rate
- Shieh and Djordjevic, OFDM for Optical Communications: extracts of section 1.4 Single-Carrier or Multicarrier Transmission: An Optical Debate
- Prasad, OFDM for Wireless Communications Systems: extracts of section 1.2 Orthogonal Frequency-Division Multiplexing and section 1.3 Concluding Remarks
- Kalat's article: The Multitone Channel
- Multi-carrier digital communications: theory and applications of OFDM / Ahmad R.S. Bahai and Burton R. Saltzberg