The USC Andrew and Erna Viterbi School of Engineering USC Signal and Image Processing Institute USC Ming Hsieh Department of Electrical Engineering University of Southern California

Technical Report USC-SIPI-232

“Analysis and Optimization of Lens-Based Optoelectronic Shuffle Networks”

by Andrew S. Miller

February 1993

We examine in detail the capabilities of the lens-based free-space optical shuffle. Specifically, we characterize the system by developing both the input-to-output and output-to-input relations for the folded and separable shuffles. We examine the physical optics limitations on the resolution capabilities of the lens-based shuffle for the case of simple, single-element lenses. Then we show how the geometry of the system may be optimized to shuffle the greatest number of channels. We show how this method may be extended to the general case of aberration corrected lenses. We then show how deviations from the ideal design affect the performance of the system in terms of crosstalk between the channels. We develop the statistical rules governing optical crosstalk in an array of channels.

Finally, we suggest a pair of ways to improve the design and performance of a shuffle-exchange network. The first is to multiplex many shuffles through a single set of shuffle optics. We show how this can be done, and what the consequences are both in terms of benefits and drawbacks. We then suggest a hybridization of two different shuffle architectures, the lens-based and holographic shuffles, which combines the desirable aspects of both systems while eliminating the undesirable traits.

To download the report in PDF format click here: USC-SIPI-232.pdf (7.0Mb)