Jin-Woo Choe

Construction of Scalable 2-D Multi-Weight Optical Orthogonal Codes for Optical CDMA Networks

In this letter, we propose combinatorial and search construction methods of 2-D multi-weight optical orthogonal codes (OOCs) with autocorrelation 0 and crosscorrelation 1, called multi-weight single or no pulse per row (MSNPR) codes. An upper bound on the size of MSNPR codes is derived and the performance of MSNPR codes is compared to those of other OOCs in terms of the bit error rate (BER) and evaluated using blocking probability. It is also demonstrated that the MSNPR codes can be flexibly constructed for different applications, providing the scalability to optical CDMA networks.

Optimal Code Weight Selection for 2-D Optical CDMA Systems

This paper presents the detection performance of the twodimensional optical CDMA network under varying conditions of the bitlevel design parameters such as the chip length (M), the number of wavelengths in use (L), and the code weight (W). In particular, the code weight that results in the minimal BER is analytically determined as W ? 1.4LM/m under the assumption of sparsely distributed random code selection, where m represents the number of active nodes in the network. The derived expression is shown to closely match the optimal code weight obtained via simulation for the cases of practical interest. The BER performance using the optimal code weight when the number of active nodes is not known (and hence estimated) is also presented and compared with the optimal performance.

A Comparative Study on the Performance of Detectors and Interference Models for OCDMA LANs

2-D OCDMA is considered to be a viable technical solution for optical LANs, and a considerable amount of research effort has been devoted to various 2-D OCDMA techniques. In this paper, we propose two new interference models for 2-D OCDMA LANs employing unipolar random codes, and derive maximum-likelihood detectors based on these interference models. The BER performance of the maximum likelihood detectors and that of other existing detectors are compared through extensive computer simulations. In addition, the complexity of high-speed implementation of the detectors is assessed, and as a result, we found that the AND detector and the maximum-likelihood detectors for the pulsebinomial and the pulse-Poisson model offer the best trade-off between the BER performance and the facility of high-speed implementation.