Eco-driving functions attract the automotive industry’s attention as they can significantly improve the energy efficiency of driving in theory, while gaining the great benefits of these functions in practice was still hindered by several remaining challenges (e.g., real-time implementation of the optimal control function, trade-off of multiple control objectives.). In this dissertation, a novel Eco-driving function named energy-optimal adaptive cruise control (EACC) is proposed, which uses model predictive control (MPC) to optimally plan the vehicle’s speed trajectory for both higher energy efficiency and better driving safety, through efficiently exploiting both road and traffic information ahead. The most suitable mathematical formulation of the MPC problem is made in this work for the optimal control of both electric vehicles and gasoline cars. To achieve the real-time implement of MPC-based EACC, a new interior-point method is developed, which can solve the convex optimization problems faster than several state-of-the-are commercial solvers according to the benchmarking results in this dissertation. To maximize the performance of the host car controlled by EACC, prediction models based on neural networks are developed to make a more accurate prediction of the preceding car’s behavior, which is provided to EACC as one control input. Furthermore, a comprehensive testing of EACC is done in this work to analyze EACC’s performance with several critical judging criteria, including energy saving, driving safety, driving comfort and car-tracking ability. Based on the testing results, it is proven that the host car controlled by EACC performs considerably better than its preceding car in various real driving conditions.weiterlesen