Chance constrained stochastic MPC with additive disturbance as robust charge-sustaining strategy for hybrid electric vehicles

This contribution presents the design of a stochastic tube-based model predictive controller for the application of energy management in hybrid electric vehicles. While previously proposed strategies commonly assume the driving profile to be perfectly known, the effects of uncertain driver behavior, i.e. uncertain driver torque demands, are neglected. In this paper, a stochastic control strategy is introduced which also expects some apriori knowledge of the driving profile but is able to react on variations in this profile. By using chance constraints on the state of charge, the strategy is capable of being charge sustaining while optimizing fuel consumption in presence of uncertainties.