A Novel Algorithm for Predictive Control of Parallel Hybrid Powertrains based on Dynamic Programming

A novel algorithm for predictive control of parallel hybrid vehicle powertrains is presented. The algorithm uses information from GPS and digital maps to schedule the use of the energy buffer along the planned route. The algorithm is based on dynamic programming and achieves close to the theoretical minimal consumption when simulated on measured drive data. For simulated routes with a topographic profile that contains large hills the fuel consumption savings compared to a competitive non predictive controller are 6%. For simulated routes with a more moderate topographic profile the savings are between 2-3% and for routes with completely flat topographic profile the savings are only between 0.5-