Evaluation and harnessing of noise in telecommunication systems

Lead Research Organisation: Aston University


Enhanced capability in digital communication is recognized as a pivotal element in a future economy, education, commercial and social activities. The role of fibre communications, currently providing for a healthy fraction of the total information traffic, will considerably increase in the near future. This will be driven by the introduction of and advances in new services such as videophony, video on demand, computer based banking, high speed computer communications and other elements of economical and social infrastructure. Light-wave fibre communication systems form the backbone of the present day high-speed data transmission links. The information is encoded and transmitted through an optical channel as a digital bit stream. One of the principal issues to be addressed by the researchers and system designers is to estimate the probability of errors (bit error rate) occurring in such bit stream due to the intrinsic noise in a transmission system. Noisy communication links need special solutions - primarily special coding techniques to minimise noise/errors in the information bit-stream. Knowledge of noise statistics of nonlinear fibre channels is crucial for the design of efficient and adequate coding techniques and practical realizations of coding devices. Therefore, novel mathematical, theoretical and numerical approaches are required to understand error statistics in order to create advanced future communication systems. Another challenging and exciting task is to find the way to use the always present intrinsic noise to our advantage. This seems like quite a paradoxical idea: for generations noise has been considered as nuisance that must be eradicated at all cost. However in certain nonlinear systems, including electronic circuits and biological sensory systems the presence of noise can enhance the detection of weak signals. The phenomenon is known as stochastic resonance and is of great interest to electronic instrumentation and telecommunications. The question arises whether this phenomenon can be used in communications and can the new devices be constructed based on this effect. In the current project we first plan to apply a novel simulation technique for studying the statistics of the signal output in optical fibres. This approach will allow us to obtain precise and explicit description of the statistics of the transmitted signal. These in turn, are crucial for determining the bit error rate and system performance. The advantage of the proposed algorithm is that it allows one to model the statistics of extremely rare events, something, which is utterly impossible to achieve with the conventional Monte Carlo techniques that are widely used at present. Therefore we expect that with this new powerful tool we will be able to calculate numerically the error probabilities, which are of the order of 1 error event per 1,000,000,000 transmitted bits. Our method will allow us to obtain these results without significant CPU time penalty, which makes the proposed approach very practical. This technique is crucial for the correct estimation of the bit error rate and, therefore, the overall system performance. We will then develop the techniques that will help to improve the quality of detection reducing the probability of mistaking logical ones for zeros and vice versa. The final and most challenging stage of the project is to propose and develop nonlinear electronic devices that make use of the phenomenon of stochastic resonance. During this stage of the project we will use the data about signal statistics collected at the initial stages of the project by Monte Carlo simulations. Currently the proposed host institution (Aston University) possesses state-of-the-art facilities for the project (including 144 node Cray XD1 supercomputer) and the proposed research is a logical development of the existing techniques and has the potential to produce a de-facto standard for modelling of noise in telecommunications.


10 25 50