Control of stochastic systems involving non Gaussian statistics

Control algorithms for stochastic uncertain nonlinear systems have been recently developed. In these methods, functional uncertainty is restricted to follow a Gaussian type density functions. This paper proposes a novel control algorithm for stochastic uncertain nonlinear systems involving non Gaussian statistics. The considered system is subjected to a non Gaussian random input and the purpose of the control input design is to make the mean of the output probability density function of the system, tracks a predefined desired output. Non Gaussian probability density functions in this paper are assumed to be unknown, therefore, estimated using mixture density networks. A simulated example is used to demonstrate the use of the proposed algorithm and encouraging results have been obtained.