Probability Density Function Control for Stochastic Nonlinear Systems using Monte Carlo Simulation

This paper presents an implementable framework of output probability density function (PDF) control for a class of stochastic nonlinear systems which are subjected to non-Gaussian noises. The statistical properties of the system outputs can be adjusted by shaping the dynamic output probability density function to track the reference stochastic distribution. However, the dynamic probability density function evolution is very difficult to obtain analytically even if the system model and the stochastic distributions of the noises are known. Motivated by Monte Carlo simulation, the dynamic probability density function can be estimated by sampling data which forms the contribution of this paper. In particular, the sampling points are generated following the stochastic distribution of the noise for each instant. These points go through the system and generate the histogram for system outputs, then the dynamic model can be established based on the dynamic histogram which reflects the randomness and the nonlinear dynamics of the investigated system. Based on the established model, the output probability density function tracking can be achieved and the simulation results and discussions show the effectiveness and benefits of the presented framework.