Improvement of predictive energy efficiency optimization using long distance horizon estimation

Energy efficiency has become an important topic in trade, transportation and environment protection. Modern electric vehicles usually still have difficulties in reaching similar travel distances as combustion engine powered vehicles. While increasing the range of electric vehicles continues to be an active field of research, it is already possible to increase the energy efficiency by applying a more energy efficient driving behavior. A forward-backward Dynamic Programming model predictive optimization approach is used to generate an energy efficient velocity and gear change trajectory. Due to the finite length of the computation horizon, common Dynamic Programming approaches sometimes have problems of choosing the optimal state at the end of the horizon. To address this problem, a method is presented that makes use of historic accumulated minimum costs to create a separate time-invariant auxiliary horizon that grows during the journey. The auxiliary horizon is used to yield a better long range estimation of the optimal terminal behavior of the optimal trajectory within the regular horizon. While the proposed method can be applied to different types of optimization problems, the focus is on the predictive energy efficiency optimization of electric vehicles.