We combine conditional state density construction with an extension of theScenario Approach for stochastic Model Predictive Control to nonlinear systemsto yield a novel particle-based formulation of stochastic nonlinearoutput-feedback Model Predictive Control. Conditional densities given noisymeasurement data are propagated via the Particle Filter as an approximateimplementation of the Bayesian Filter. This enables a particle-basedrepresentation of the conditional state density, or information state, whichnaturally merges with scenario generation from the current system state. Thisapproach attempts to address the computational tractability questions ofgeneral nonlinear stochastic optimal control. The Particle Filter and theScenario Approach are shown to be fully compatible and -- based on the time-and measurement-update stages of the Particle Filter -- incorporated into theoptimization over future control sequences. A numerical example is presentedand examined for the dependence of solution and computational burden on thesampling configurations of the densities, scenario generation and theoptimization horizon.
展开▼
