The Ultimate Limit of Optical Fibers to Carry Information

Prof. René-Jean Essiambre
Bell Labs, Alcatel-Lucent

Approx. 1 hour / 9:30-10:30

Abstract:

A network of fused-silica single-mode optical fibers forms the backbone of the worldwide communication infrastructure. The maximum rate of transmission of information over such fibers has increased by four orders of magnitude over the last three decades. A question naturally arises: is there a fundamental limit to the maximum achievable transmission rate, or capacity, of single-mode fibers?

After a brief historical perspective on guided-wave transmission and a review of Shannon’s information theory, the optical Kerr effect that characterizes the nonlinear behaviors observed in optical fibers will be discussed. It will be followed by a description of the ‘optical fiber channel’ pertinent to optical networks and of strategies devised to maximize the transmission of information over fibers, including the use of temporal solitons and nonlinear propagation reversal. Results of calculations of the nonlinear Shannon capacity limit will be presented and it will be shown that we have now reached 50% of this capacity limit in record experimental demonstrations. The consequences for the internet backbone of being so close to the nonlinear Shannon capacity limit will be outlined. The use of multicore, multimode and photonic bandgap fibers as a way to increase the capacity per fiber strand beyond the nonlinear Shannon capacity limit of single-mode fibers will be presented. Finally, the nonlinear propagation equations in fibers supporting multiple spatial modes will be introduced.

Biography:

René-Jean Essiambre studied at McGill University in Montréal and Université Laval in Québec City, Canada from which he received a Ph.D. degree in Physics (Optics) in 1994. From 1995 to 1997, he was at The Institute of Optics of the University of Rochester, Rochester, New York, USA. Since 1997, he has been at Bell Laboratories, Alcatel-Lucent, Holmdel, New Jersey, USA. His early research focused on chaos theory, optical switching, optical solitons, high-power and mode-locked fiber lasers. His current research interests include nonlinear dynamics in optical fibers and information theory applied to fiber-optic communication systems as well as in space-division multiplexing in multimode and multicore fibers for high-capacity transmission. He is the author and coauthor of more than 150 scientific publications and several book chapters. He has served on or chaired many conference subcommittees including ECOC, OFC, CLEO, and LEOS and is the Program Co-Chair of CLEO: Science & Innovations 2012 and General Chair for 2014. Dr. Essiambre is a Fellow of both the Optical Society of America (OSA) and the Institute of Electrical and Electronics Engineers (IEEE), is the recipient of the 2005 OSA Engineering Excellence Award and is a Distinguished Member of Technical Staff (DMTS) at Bell Laboratories. He is currently a Rudolf Diesel Fellow of the Institute of Advanced Studies (IAS) of Technical University of Munich (TUM). He holds 19 patents.

Zero-Error Compression of Markov Chains

Dr. Bernhard Geiger
LNT, TUM

Approx. 20mins / 10:35-10:55

Abstract:

In this work, we apply methods from fixed-length zero-error source coding with side information to the compression of sparse Markov chains. Taking the previous state of the Markov chain as side information, we show that the required number of symbols to encode the current state can be less than the alphabet size. While encoding is performed by applying a non-injective function (that is related to the cliques of a particular graph) coordinate-wise to the sequence of states, successive decoding recovers this sequence with zero error. We finally show that block coding is asymptotically optimal in the sense that it does not require more symbols than there are possible sequences.

Biography:

Bernhard began working towards an information-theoretic system theory as a member of the Signal Processing and Speech Communication Lab at Graz University of Technology, Austria, in 2010. He received his PhD in 2014 and is currently a senior researcher at LNT. His main research interest are model complexity reduction techniques for Markov chains, such as quantization of transition probability matrices, state space aggregation, and the phenomenon of lumpability.

Coding schemes for discrete memoryless multicast networks with feedback

Dr. Youlong Wu
LNT, TUM

Approx. 20mins / 11:00-11:20

Abstract:

In this work, we propose coding schemes for discrete memoryless multicast networks with rate-limited feedback from the receivers and relays to the transmitter. These coding schemes are based on block-Markov coding, joint backward decoding and hybrid relaying strategy. In each block, the receivers and relays compress their channel outputs and send the compression indices to the transmitter through the feedback links. In the next block, after obtaining the compression indices, the transmitter sends them together with the source message. Each receiver uses backward decoding to jointly decode the source message and all compression indices. Our coding schemes generalize Gabbai and Bross's results for the single relay channel with partial feedback, where they proposed coding schemes based on the restricted decoding and deterministic partitioning. It is shown that for the single relay channel with relay-transmitter feedback, our coding schemes improve the noisy network coding scheme by Lim, Y-H Kim and El Gamal, the distributed decode-forward coding scheme by Lim, Kim and Y-H Kim and all known lower bounds on the achievable rate in the absence of feedback. Furthermore, motivated by the feedback coding schemes, we propose a new coding scheme for discrete memoryless multicast networks without feedback, which can still potentially improve the noisy network coding and distributed decode-forward coding schemes.

Biography:

Youlong Wu obtained his B.S. degree in electrical engineering from Wuhan University, Wuhan, China, in 2007. He received the M.S. degree in electrical engineering from Shanghai Jiaotong University, Shanghai, China, in 2011. In 2014, he received the Ph.D. degree at Telecom ParisTech, in Paris, France. Since December 2014, he has been a postdoc at the Institute for Communication Engineering, Technische Universität München (TUM), Munich, Germany. His research interests include information theory and wireless communication.

Signal-Shaping Channel-Coding Separation

Dr. Georg Böcherer
LNT, TUM

Approx. 30mins / 11:25-11:55

Abstract:

Recently, in http://arxiv.org/abs/1502.02733, a practical shaping scheme for bandwidth-efficient and rate-matched transmission over the AWGN channel was proposed. In this talk, the scheme is shortly explained. Two channel coding theorems are then presented, which state that the scheme achieves the capacity of the finite input AWGN channel. The limitations of the presented channel coding theorems are pointed out and it is discussed to which extend signal-shaping and channel-coding are separable.

Biography:

Georg Böcherer obtained his MSc degree in Electrical Engineering and Information Technology from the ETH Zürich. From 2007 to 2012 he worked towards his PhD degree at the Institute of Theoretical Information Technology at RWTH Aachen University. His PhD thesis on probabilistic shaping received the E-Plus Dissertation Award. His work on coding for the ICT Cubes received the best paper award at ISWCS 2011. Since April 2012, he is with the Institute for Communications Engineering at Technische Universität München. His current research interest is the design of capacity-achieving transceivers.