The Digital Random Signal Simulator

Abstract

There is introduced a universal simple digital simulator of the random signal with the specified arbitrary two-dimensional probability distribution of its discrete values based on the Markov model. It allows us to generate an unlimited sequence of samples with the required probabilistic and correlation properties of adjacent random or pseudo-random values of the simulated random process. By applying the simplified version of the simulator, the independent signal samples obeying the specifying arbitrary one-dimensional probability distribution can be formed. The simulator design involves the technique of determining the parameters of the Markov model of the random signal values at the two time moments. The Markov model can be created based on either the specified two-dimensional probability density or the experimental sample of the simulated random process. There is presented the algorithm for signal sample generating that provides a high sampling frequency. Its hardware implementation by means of microprocessor devices or field-programmable gate arrays is also in focus. The computational algorithm procedure does not require complex mathematical transformations of the signal samples and is implemented using an inexpensive circuitry. Changing the statistical properties of the simulated random processes is achieved by either rebooting the storage device with the pre-formed data array or switching the memory pages in which the necessary arrays are stored. In order to demonstrate the performance of the simulator and its high efficiency, the properties of the simulator, its probabilistic and correlation characteristics are studied. It is shown that a high accuracy of coincidence is ensured between the two-dimensional probability distribution of the selected model and the histogram based on the generated sequence of samples of the random signal. #COMESYSO1120