Channel Capacity in Quantum-Noise-Limited Optical Communication Systems

Matthew Thill, Samuel J. Dolinar, D. Divsalar

E. Desurvire

We analyze the information-capacity limitations of the optical communication channel, as determined by noise accumulation from optical amplification and nonlinear wave-mixing. We review the concepts of signal-to-noise ratio and entropy for binary-coded and continuous communications, leading to a definition of ultimate capacity for the optically-amplified channel. A unified quantum model, describing both amplification and nonlinearity limitations, makes possible to determine the power transmission window within which the channel capacity can be maximized.

Victor A. Vilnrotter, Chi-Wung Lau

The fundamental performance limits and channel capacity of optical communications systems operating over the free space channel will be examined using quantum detection theory. The performance of the optimum quantum receiver for on-off keying (OOK) and optical binary phase shift keying (BPSK) is first examined as a pure state (no noise) problem. The classical capacity of the binary symmetric channel for these two modulation schemes is then evaluated for the optimum quantum receiver by making use of the concept of quantum measurement states. The performance of M-ary pulse position modulation, which requires a product state representation, is evaluated along with the performance of certain 'dense signal sets.' Performance comparisons with classical techniques shows over 5 dB improvement in some cases when quantum detection is employed. As a further application of the quantum detection theory, the capacity of the binary channel with on-off keyed modulation and quantum detection is evaluated, and shown to exceed the capacity obtained with classical photon counting.

Monte Ross

Konrad Banaszek

Richard E. Slusher, Bernard Yurke

Coherent lightwave communications systems are approaching a limit where the error rates and channel capacities are limited by the quantum properties of light. This is often referred to as the shot-noise limit. If ideal laser light is used in the system, there is no way to avoid this limit. However, new states of the light field called squeezed states have recently been developed that allow an improvement in error rates below the shot-noise limit. Squeezed light concepts and recent experiments are reviewed with emphasis on aspects important to coherent communications. It is shown that channel capacity can be improved using squeezed light by only a factor of two. Larger improvements are in principle possible for error rates, e.g. a factor of three reduction in the number of required photons per bit for a 10/sup -9/ bit error rate. An example of a recent high-performance system is described where optical losses and electronic noise reduce the improvement expected using squeezed light to the 10-20% level. It is concluded that squeezed light only offers significant improvement in bit error rates for very-high-efficiency systems.>