马尔可夫链的混合动力汽车行驶工况预测研究及概要.pdfVIP

学兔兔 汽 车 工 程 2014年(第36卷)第l0期 Automotive Engineering 2014228 基于马尔可夫链的混合动力汽车行驶工况预测研究木 张 昕 ，王松涛 ，张 欣 。田 毅。 (1．北京交通大学机械与电子控制工程学院，北京 100044；2．酒泉卫星发射中心铁路管理处，酒泉 732750) [摘要] 针对目前混合动力汽车行驶环境适应性的问题，提出一种适应于随机多变的交通状况和基于实时行 驶工况预测的能量管理策略。该策略采用一阶齐次马尔可夫链状态转移概率矩阵，对主干道和快速路建立了行驶 工况特征参数预测模型，并通过对预测准确度的对比分析，确定了滚动时间窗和预测视距的最佳值。研究结果表 明，与传统基于既定循环工况的整车能量管理策略相比，基于工况预测的能量管理策略具有更好的燃油经济性，为 新能源汽车能量管理实时控制提供了一种新方法。 关键词：混合动力汽车；行驶工况；预测模型；马尔可夫链 A Research on Driving Condition Prediction for HEVs Based on Markov Chain Zhang Xin ，Wang Songtao ，Zhang Xin Tian Yi 1．School ofMechanical Electric and Control Engineering，Beijing Jiaotong University，Being 100044； 2．Railway Management Bureau of JSLC，Jiuquan 732750 [Abstract] Focusing on the issue of HEV adaptability to driving environment，a novel energy management strategy adapting to random and varying traffic situation is proposed based on the real time driving condition predic— tion．The strategy uses the state transition probability matrix of first—order homogeneous Markov chain to set up a driving condition characteristic parameters prediction model for both trunk roads and expressways，and the optimal values of rolling time window and preview distance are determined by comparative analysis on prediction accuracy． The results of research indicate that the energy management strategy based on driving condition prediction model can achieve better fuel economy than that with conventional energy management strategy based on preset driving cycle， providing a new way of real-time energy control for new energy vehicles． Keywords：HEV；driving conditions；pr