FPGA-based fractional Brownian motion signal-patterns for optial paket generation

Signal-pattern generation with different distributions and statistical parameters is an important issue for optical communications performance evaluation. This generation must include the signals´ scaling or time-dependence degree representation. This paper proposes a new field-programmable gate array (FPGA)-based architecture to generate signals with fractional Brownian motion (fBm) patterns. The proposed implementation enables real-time performance experiments for optical communications using an Altera´s Stratix Device Family. Uniform and a Gaussian sample generations were also implemented in this paper. The FPGA blocks were validated considering a real H.263 encoded video sequence. The experiments show that the proposed fBm block architecture well-fitted the statistical parameters of the real video sequence and the fBm pattern. The proposed solution also presents an interesting trade-off between FPGA device resources usage and a maximum clock frequency.