Predictive energy management for hybrid vehicle
Résumé
Hybrid vehicle control strategies are algorithms devoted to the energy management. At each sampling time they choose the powertrain operating point in order to minimize a criterion, usually the fuel consumption. In simulation, an optimization algorithm can be derived from the minimum principle. For real time control, a Model Predictive Control scheme can be used but it requires the prediction of the future driving conditions. Their time ordered prediction is very difficult. Moreover, if the optimal costate is constant, only the prediction of their distribution is sufficient and allows deriving a real time control strategy. Experimental results are provided to illustrate the benefits of this approach.