Weighted-Probability Random Number Generator for PLC Channel Transfer Function Generation

Abstract

The concept of randomization shows many realistic scenarios and has enormous applications in everyday life. However, existing random number generators have poor adaptability when representing some of the actual occurrences in natural environment, such as in biology, mobile communications, and even electrical behavior. For PLC channel transfer function generation, biased sampling of the channel is more applicable. This paper proposes a weighted-probability method to generate random numbers to emulate the real-world variations of a power line communications channel. Existing PLC channel emulator can mimic the random behavior of the channel, altered by several parameters such as cable length, cable type, presence of loads, and ejecting different noises. However, the scenario of integrating the effect of changing the state of the loads and channel variation due to time is not well represented in the normal random channel generation. The best concept for this scenario is the weighted random number generator (WRNG). This work developed an algorithm in Matlab and VHDL, implemented it in FPGA and integrated it with an existing PLC channel emulator. The emulator is implemented using Kintex-7 FPGA. In this study, ten different weights corresponding to the probability of occurrence of the actual household appliances and eighty transfer functions were used as inputs to the program. The randomness and periodicity of the WRNG are tested and passed the Runs test. Autocorrelation and histograms are used to determine the correct implementation of the weights.