驾驶风格自学习的智能电动汽车预测节能控制研究

The energy-saving predictive control with intelligent traffic information for electric vehicles greatly improves the energy utilization efficiency, and is a research hotspot in the field of automotive control systems. However, simply pursuing energy-saving goals often leads to driving feelings of “uncomfortableness” and “slow power response”, which is not conducive to the application of energy-saving technology. The project aims to improve the energy utilization efficiency while taking into account the individual driving style. The research contents include system modeling, algorithm design, driving style self-learning, algorithm implementation, real vehicle verification, etc. The project is expected to achieve the following goals: 1) establishing an energy consumption model of electric vehicles for energy-saving predictive control; 2) proposing an energy-saving predictive control method for intelligent electric vehicles; 3) providing a self-learning energy-saving predictive control method by learning individual driving style online; 4) designing a fast solution algorithm for nonlinear energy-saving predictive control; 5) developing a test platform to evaluate the proposed energy-saving predictive control method. Through these researches, the bad driving feelings of “uncomfortableness” and “slow power response” will be eliminated. This project will not only provide key technical support for the development of energy-saving predictive control systems, but also promote the development of energy-saving technologies in the field of intelligent electric vehicles.

智能交通环境下的电动汽车预测节能控制极大地提高了能源利用效率，是汽车智能控制领域的研究热点。然而，单纯地追求节能目标，往往会带来“节能不舒适”和“节能动力响应慢”的驾驶感受，不利于技术的应用推广。为了提高能源利用效率的同时兼顾驾驶风格个性化，本项目围绕智能电动汽车预测节能控制理论和技术问题，从系统建模、算法设计、算法快速实现、实车验证等多方面开展深入研究。预期建立面向预测节能控制的电动汽车能耗模型；提出智能电动汽车预测节能控制方法；通过学习归纳个人跟车风格，实现驾驶风格自学习的预测节能控制技术；设计非线性预测节能控制快速求解算法；开发智能汽车预测节能控制试验平台，完成系统的主客观性能验证，最终解决系统“节能不舒适”和“节能动力响应慢”的驾驶感受差问题。项目将为高精地图和智能交通背景下电动汽车预测节能控制系统开发提供基础理论和关键技术支撑，进而推动我国智能电动汽车综合节能技术的发展。

本项目围绕智能电动汽车驾驶风格自学习的预测节能控制问题，分析了复杂道路交通环境下的车辆能耗机理，建立了面向预测节能控制的电动汽车能耗模型；构建了车辆非线性纵向滚动优化控制问题，提出了智能电动汽车预测节能控制方法；应用极大值原理，设计了非线性预测节能控制的快速求解算法；针对驾驶员风格辨识问题，提出了基于驾驶工况的驾驶风格辨识方法，发明了驾驶风格自学习的智能电动汽车预测节能控制技术，设计了基于个人跟车风格学习的自适应优化跟车控制系统，该系统在实车平台上得到了验证，实验结果显示：相比于量产的自适应巡航控制器，设计的优化控制器可以自适应驾驶员的风格，其跟车行为与特定驾驶员更相似。