Predictive Energy Management on Multi-core Systems

To meet climate agreements, electric vehicles will become very important in the next few years. One class of electric vehicles are Fully Electric Vehicles (FEVs) which are purely powered by electric energy. At the moment the driving range of FEVs is very limited compared to vehicles driven by internal combustion engines. To contribute to increasing the driving range, a comprehensive energy management and enhanced control algorithms like the Model Predictive Controller (MPC) are introduced. The MPC presented in this book chapter is a first approach of solving a reference speed tracking problem on a multi-core platform in real-time. First a high performance off-board system generates an energy and time optimal reference speed profile. This optimization problem is solved by a dynamic programming approach with respect to speed limits, track profile and cloud-sourced data. Afterwards, the reference speed profile is provided to a MPC that is implemented on the vehicle's on-board system. When utilizing a multi-core computing platform, the on-board system provides information about current speed, position and the driver's torque demand, which is made available by the MPC. To enable the concept of a comprehensive energy management, a revised information and communications technology reference architecture is presented, as well as an overview about the available hardware and toolchain is given.