Course #09

Turbo Codes and Low-Density Parity-Check Codes: Analysis, Design, Iterative Decoding and Applications

New date to be decided. 
Available as a corporate-exclusive / tailored course at your site.
Please contact Ms Elisabet Larsson for an offer: 
cei@cei.se   Phone: +46-13-100 730   Fax: +46-13-100 731

INSTRUCTORS
Professor Sergio Benedetto, Politecnico di Torino, Italy 
Professor Guido Montorsi, Politecnico di Torino, Italy

TECHNOLOGY FOCUS
"Turbo codes", introduced in 1993, and low-density parity-check (LDPC) codes, introduced by Gallager in the sixties and revisited after turbo codes invention, are the most exciting and important development in coding theory in many years. Researchers around the world have been able to extend the basic idea to other forms of code concatenations, with various applications for transmission over fading channels, band-limited satellite channels, and channel with inter-symbol interference. 

After slightly more than a decade from their discovery, turbo and LDPC codes have been accepted as coding standard for 3G wireless communications systems, like CDMA2000 and UMTS, for satellite and deep space applications as the new CCSDS telemetry channel coding standard, for the new digital video broadcasting by satellite DVBS-2, and many others. Both classes of codes rely on the application of soft, decentralized decoding algorithm, like the BCJR for turbo codes and the "message passing" for LDPC codes.

Beyond the field of channel coding, the "turbo" principle, i.e., the class of soft, decentralised algorithms that drastically improve communication system performance by exchanging soft information among the receiver parts, has been proposed for several other applications, like multi-user interference cancellation, continuous-phase modulation, joint demodulation/decoding, etc. 

COURSE CONTENT
This 5-day course provides an understanding of the principles governing the turbo and LDPC codes behaviour, analytical tools to evaluate the maximum likelihood performance, and design rules, including the tool of EXIT charts, to optimise their performance. Explanation of the maximum-a-posteriori algorithms, which form the core of the iterative decoding algorithms, extensive analytical and simulation results, and a comparative analysis of the implementation complexity are also presented. Emphasis in the design will be given to the versatility of the envisaged schemes, in terms of code word length, code rate and spectral efficiency. So, the design will encompass bit-interleaved turbo trellis coded modulations and flexible, prunable interleavers.

The important topic of very high-speed implementation, requiring a high degree of parallelism in the decoder architecture, will be treated in detail.

The final part of the course will be devoted to describe a broad range of applications, including the UMTS, CCSDS and DVBS-2 coding standards. The general "turbo" principle will be applied to channel equalization, and multi-user interference cancellation. High-speed software written in C++ implementing the iterative decoding algorithms for parallel and serial concatenations will be explained and made available to the course participants.

Monday 

Theoretical Limits to Coding Gain from Information Theory - Performance Analysis and Design of Turbo Codes and LDPC codes 

• The Information-theoretical Limits of Coding: The coding theorem and the role of capacity and cutoff rate
• Analysis of Concatenated Codes with Uniform Interleaver and LDPC Codes
• Design Criteria for Concatenated Codes and LDPC Codes
• Construction of Interleavers and Parity- check Matrices

Tuesday 

The Iterative Decoding Algorithm and the Soft-Input Soft-Output (SISO) Basic Modules 

• Iterative Decoding of Turbo and LDPC Codes 
• The Basic Soft-input Soft-output Modules for Turbo and LDPC Codes 
• Optimum and Suboptimum Algorithms: Performance/complexity trade-offs

Wednesday 

Iterative Decoding of Serially Concatenated Codes - Multiple Concatenations 

• SW Description and a Numerical, Step-by-step Example of the Decoding Agorithm for Turbo Codes
• SW Description and a Numerical, Step-by-step Example of the Decoding Algorithm for LDPC Codes
• Examples of Simulation Performance
• The Design Technique Based on EXIT Charts and Density Evolution 

Thursday 

Design of Versatile, High-speed Turbo and LDPC Co-decoders for Bandwidth and Power Efficient Systems

• An Overview of Trellis-coded Modulation
• Capacity of Two-dimensional Signal Sets
• Bit-interleaved Trellis Coded Modulation Based on Turbo and 
  LDPC Codes
• Design of Flexible Interleavers and Parity-check Matrices
• Puncturing Strategies
• Parallel Architectures for High-speed Decoders and Their Implementation

Friday 

Applications of Turbo and LDPC Codes - The Turbo Principle at Large 

• Standard Codes for UMTS, CCSDS, DVBS-2, DVB-RCS, 802-11 Class
• Turbo Equalization 
• Turbo Multiuser Detection
• Turbo Space-time Codes and MIMO Systems

Regular Course Fee, open events: N/A

Early Registration Course Fee, open events: N/A
This applies to firm registrations received 2 months before course start. 

University Student and Faculty Rate:
Two university participants are welcome to attend for one course fee if payment is to be made from university funds.

Deliverables:
The course fee covers tuition, course material, and the day conference packages (morning/afternoon refreshments, lunches etc.) paid on your behalf to the course venue. Accommodation is not included.