Hybrid analytical/computational methods for the identification of errors in lightwave systems

Northwestern University, Evanston IL

Investigators

Abstract

The development of high-speed optical systems to transmit and process information is a major technological achievement of the late twentieth century. The goal of this project is to develop new hybrid analytical/computational techniques capable of identifying the errors that limit the performance of such systems, and the probabilities with which they occur. The first component of the project is the application of the singular value decomposition to the equations governing the combination of optical pulse propagation and signal detection at the end of transmission that converts light into electrical energy, which will determine the perturbations that produce the largest changes at the output. The second component is the application of the cross-entropy method, an adaptive variance reduction technique which assesses the relative importance, in terms of probability, of each type of perturbation. Experimental testing and laboratory or field measurements of optical systems can be quite costly and require months of time. A goal of this project is to produce simulation methods that can provide, in an efficent manner, detailed information about system performance, thus reducing the need for testing. The methods will be used to model both high-speed optical communication systems and new classes of extremely stable short-pulse lasers. These lasers provide ultra-precise frequency references for communication, radar, and remote chemical sensing, and are an essential component in optical atomic clocks, next generation timekeeping devices that are predicted to have accuracies several orders of magnitude better than current atomic clocks. Improved accuracies in global position systems and other technologies are expected to result from these devices.