Expected Time-Optimal Control: a Particle Model Predictive Control-based Approach via Sequential Convex Programming

In this paper, we consider the problem of minimum-time optimal control for a stochastic system with initial state uncertainties and propose a sequential convex programming (SCP) framework. We seek to minimize the expected terminal time, which is an essential capability for planetary rover missions to carry out scientific missions efficiently under uncertain environments. This proposed framework utilizes a partial model predictive control with a consensus control horizon and the sum-of-norm objective to make the aforementioned stochastic optimal control problem numerically tractable for SCP. To this end, our numerical simulations for linear and nonlinear systems indicate that our proposed approach leads to more robust trajectories than the deterministic approach.