Implementation of an Optimal Control Like Energy Management for Hybrid Vehicles based on Driving Profiles 2014-01-1903

In this paper an energy management is proposed which is optimal to certain driving scenarios which can be clustered into freeway, rural and urban situations. This strategy is non-predictive but uses information about the current driving situation provided by modern navigation systems to identify the current road type. Based on this information a set of simplified optimal control problems are solved offline via an indirect shooting algorithm. By relaxation of the problem formulation, the solutions of these optimal control problems can be stored into easily implementable maps. The energy management control is then determined from these maps during vehicle operation using the current road type, the vehicle speed and the required wheel-torque. The strategy is implemented in a dSPACE MicroAutoBox and validated on a near mass-production vehicle. The proposed methodology has shown fuel savings on a real world drive cycle. Additionally, robustness aspects have been considered in a MATLAB/Simulink based simulation environment. The proposed solution of the energy management problem is proven to be real-time applicable and very robust against driver's influence.