Multiuser network coding based on DFT matrices design

Abstract

Cooperative communications are candidates for emerging distributed Network designs. In this paper, a new design of distributed network coding is introduced. The proposed design are based on algebraic constructions in Finite Fields using the Discrete Fourier Transform (DFT). These set of codes provides an alternative to Generalized Dynamic Network Codes (GDNC) and are based on the Hamming distance metric. Using this metric, the codes have maximum diversity and scales over large networks. In our work, we show that a compromise on the Field Size becomes necessary for enforcing the distributed Network codes over large networks. While the diversity offered by these codes are optimum, a unique feature of these codes are the ease of generation and scalability to larger network sizes. Results are provided and compared with theoretical evaluations.