智能汽车人机并行控制冲突机理与协同共驾关键技术研究

Human-machine co-piloting system for intelligent vehicles, which could achieve the deep integration of human and machine intelligence and complementary advantages, has become a focus and frontier technology in current intelligent vehicle research. However, theory and technology system of the human-machine co-piloting system are not yet mature. The objective of this project is to investigate and clarify human-machine parallel control conflict mechanism and key technology of cooperative co-piloting for intelligent vehicles, in driver-vehicle-environment (DVE) coupling system, through combining vehicle engineering and bionic engineering. According to the biological coupling theory, the coupling mechanism of the DVE system is to be analyzed and the human-machine parallel control interactive and conflict mechanism are to be explored. The bionic intelligent algorithm is adopted to study the characteristics and dynamic evolution mechanism of personalized driving behavior， build the symbolization and identification method of drivers’ personality characteristics. Intelligent vehicle personalized safety field considering complex and uncertain constraints is to be established based on multi-source information perception, uncertainty reasoning of artificial intelligence, etc. On the basis of bionic optimality principle, human-machine parallel control dynamic multi-objective collaborative optimized strategy is to be designed and an integrated human-machine cooperative co-piloting system is to be built. Within the implementation of this project, a set of new bionic design theory and method for human-machine cooperative co-piloting system will be explored to achieve high-quality collaboration of drivers and machine intelligence, which will greatly improve the humanization and individuation of intelligent vehicles and greatly optimize the performance of DVE system.

智能汽车人机协同共驾系统能够实现驾驶人与机器智能的深度融合和优势互补，是当前智能汽车研究领域的热点和前沿，但其理论和技术体系尚不完善。本项目拟通过仿生工程与车辆工程的学科交叉，以人-车-环境耦合系统为研究对象，研究掌握智能汽车人机并行控制冲突机理与协同共驾关键技术。通过生物耦合理论对人-车-环境系统进行耦合机理分析，探索人机并行控制交互机制与冲突机理；采用仿生智能算法研究驾驶人个性化驾驶行为特征规律与动态演化机理，建立驾驶人个性特质表征与辨识方法；采用多源信息感知、人工智能不确定性知识推理等方法构建考虑复杂不确定性约束的智能汽车个性化安全场，根据最佳化仿生原则，设计人机并行控制动态多目标协同优化策略，建立完整的人机协同共驾系统。本项目的开展，将探索出一套全新的人机协同共驾系统仿生设计理论与方法，实现驾驶人与机器智能的优质协同，使智能汽车更加人性化、个性化，全面优化人-车-环境闭环系统性能。

智能汽车人机协同共驾系统能够实现驾驶人与机器智能的深度融合和优势互补，是当前智能汽车研究领域的热点和前沿。本项目秉持“以人为本”的核心理念，以人-车-环境耦合系统为研究对象，研究掌握了智能汽车人机并行控制冲突机理与协同共驾关键技术：（1）采用生物耦合理论和深度数据挖掘方法对人-车-环境耦合系统进行了系统分析和抽象描述，探索了驾驶人驾驶与机器智能驾驶之间的冲突机理；（2）提出以驾驶状态、驾驶技能和驾驶习性三个个性化物理量作为对驾驶人个性特质的表征，采用多种人工智能算法建立了驾驶人个性特质多维辨识方法；（3）建立了智能汽车人机协同共驾系统控制架构，构建了综合考虑人-车-环境复杂不确定性约束的智能汽车个性化安全场，自主设计了多种个性化智能汽车人机协同共驾策略；（4）进行了系统仿真分析、驾驶人在环测试与实车测试，结果表明，本项目所设计的智能汽车人机协同共驾策略能够实现智能汽车的人性化、个性化控制，全面提升人机并行控制下智能汽车的安全性、舒适性等综合性能。通过本项目的研究，明晰了驾驶人个性化驾驶行为的特征规律和表征方法，形成了智能汽车人机协同共驾系统设计理论和方法体系，实现了驾驶人与机器智能的深度融合与优势互补。项目运行期间，在IEEE ITS、Transportation Research Part C、《汽车工程》等行业顶级期刊发表高水平学术论文18篇，已录用学术论文5篇；申请发明专利12项，其中授权发明专利8项；培养博、硕士研究生11名，荣获吉林省优秀硕士学位论文1篇，吉林大学优秀硕士学位论文3篇；依托本项目，团队先后荣获中国汽车工业科技进步奖一等奖、中国生产力促进奖（创新发展）一等奖；项目负责人入选2019年教育部青年长江学者，并先后荣获第十六届吉林省青年科技奖和2021年中国汽车工业优秀青年科技人才奖。