基于机器学习的汽车协同式自适应巡航控制机理研究

项目名称： 基于机器学习的汽车协同式自适应巡航控制机理研究

项目编号： No.51205154

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 机械工程学科

项目作者： 张晋东

作者单位： 吉林大学

项目金额： 25万元

中文摘要： 在车车通信环境下，汽车协同式自适应巡航控制是目前汽车先进智能电控技术研究的热点和前沿，是实现汽车自主驾驶的关键技术，主要是可以在汽车间相互传递行驶状况和驾驶意图信息，从全面与环境交互多车协同保证行驶性能出发，实施对汽车驱动、制动和转向的智能控制。本项目拟运用机器学习理论研究多车行驶环境感知级-状态预测级-性能评价级的多级数据信息融合机制,以此建立一套动态决策优化的汽车协同式自适应巡航控制算法：对汽车进行Agent模拟抽象，建立能够维护各个意识理性平衡的意识模型；研究环境信息部分已知或未知情况下信息感知行为，建立多车与行驶环境交互的多数据级算法；研究汽车对驾驶意图和汽车下一个行驶状态作出预测，建立状态预测级算法；研究多控制目标的驾驶安全性、多车稳定性和舒适性行为，建立协同决策的性能评价级算法。本项目的研究将为我国汽车智能控制技术的自主研发和实现跨越式发展提供良好的理论和技术支撑。

中文关键词： 汽车巡航控制；汽车协同式控制；汽车自适应控制；机器学习；车车通信

英文摘要： In V2V communication environment, vehicle cooperative adaptive cruise control is the hot and frontier of researching on vehicle advanced intelligent electric control technology currently. It is the key technology of implementation vehicle autonomous driving. It is mainly can mutually send driving status and driver intent in vehicles, for guarantee driving performance by full interactive with environment and vehicles cooperative way, to implement intelligent control to driving, braking and steering of vehicle. This project aims to use machine learning theory to research vehicle "environment aware"-"state forecast"-"performance evaluation" of multilevel data fusion mechanism. To build a set of vehicle cooperative adaptive cruise control algorithms of dynamic optimization: To achieve vehicle Agent simulation abstract, establishing models of consciousness to preserve all sense of rational balance; To research information perceived behavioral in some known or unknown environmental information, establishing multilevel algorithms of vehicles interaction with driving environment; Research vehicle making predictions about the driver intent and the next motion state of vehicle, building algorithm for forecast level; Research on the multiple control objectives driving security, vehicles stability, and comfort behavior, bu

英文关键词： Vehicle Cruise Control；Vehicle Cooperative Control；Vehicle Adaptive Control；Machine Learning；Vehicle to Vehicle Communication