考虑驾驶意图的汽车碰撞危险态势辨识与避碰决策研究
发布时间：2018-11-26 10:03
【摘要】：道路交通事故的频繁发生严重影响了人们的生命财产安全和道路交通通行效率。汽车碰撞事故是道路交通事故中最主要的一种事故形态,如何有效的防止汽车碰撞事故的发生,提高道路车辆的行驶安全性一直是汽车安全辅助驾驶领域众多学者面临的共同问题。在实际道路交通环境中,驾驶人决策意图/操控行为、汽车运动状态、道路交通环境(路面湿滑程度、道路线型等)多种因素会导致汽车碰撞事故的发生。国内外学者对汽车智能避碰技术开展了深入的研究,但迄今为止针对复杂道路交通环境下的行车安全状态辨识方法以及动态道路交通环境下的汽车避碰决策仍然存在许多有待解决的问题:如何将驾驶人因素与动态道路交通环境相结合,综合评估驾驶意图/驾驶行为、汽车运动状态以及车外环境对行车安全性的影响;如何解决“人-车-路”协同感知环境下行车安全相关多源异构信息的处理与融合问题;如何构建复杂道路交通环境下的汽车碰撞危险态势辨识模型;如何考虑道路环境和车辆动力学特性对汽车运动状态的约束问题来规划合理的避碰路径。本文的研究工作主要依托国家自然科学基金项目“基于驾驶意图与动态环境的汽车碰撞危险态势辨识研究”、“车联网环境下多意图跟踪的车辆碰撞预警方法研究”和国家863计划项目“车路交互式行车安全系统关键技术”。针对现有汽车避碰技术的不足之处,本文综合考虑了驾驶意图和动态道路交通环境对行车安全性的影响,研究了复杂道路交通环境下的汽车碰撞危险态势辨识方法和避碰路径规划方法。本文主要研究工作如下:1.对典型驾驶意图下(自由行驶、跟驰行驶、换道行驶)的驾驶人操控行为特点进行了深入研究。基于线性最优二次型理论与方法建立了驾驶意图导向下的驾驶行为预测模型,通过求解模型效用函数来量化表示驾驶人对汽车运动状态的期望操控方式。同时,通过调整模型参数来适应驾驶人习惯或道路交通环境导致的驾驶操控行为差异性。2.综合考虑驾驶意图和汽车运动状态提出了汽车行驶轨迹的短时预测方法。通过将预测的驾驶操控量序列加入到运动学模型(CA模型),并基于Kalman估计算法和改进的CA模型来预测汽车在下一时刻的运动状态。在此基础上,考虑邻域范围内的多车行车意图和运动状态,预测车辆间的行驶轨迹是否存在冲突,并定义了冲突风险指数来评估车辆间的冲突风险程度。3.对“人-车-路”多因素影响下的汽车碰撞危险态势辨识方法进行了研究。综合考虑了驾驶意图、车间时距、路面湿滑程度等因素对行车安全状态的影响,并应用态势感知理论建立了汽车碰撞危险态势辨识的模型结构。结合变精度粗糙集方法形成了汽车碰撞危险态势评估规则,再通过属性加权相似度量方法比较当前行车状态与决策规则库中条件属性的相似程度,进而得出汽车碰撞危险态势的评估结果。4.综合考虑驾驶意图和动态交通环境对行车安全性的影响,以及道路边界等条件对汽车运动状态的约束,提出了改进的道路势场模型来规划汽车避碰路径。根据被控汽车与道路障碍物间的相对运动状态以及道路车辆的行驶意图,结合栅格法对传统人工势场模型进行改进,确保被控汽车在动态道路交通环境下沿规划的避碰路径行驶时,能有效避免碰撞事故的发生。本文针对复杂道路交通环境下汽车智能避碰系统涉及的关键理论、方法与技术问题进行了深入研究。一方面,将驾驶意图引入到汽车碰撞危险态势辨识研究中,提高了汽车冲突危险态势辨识的准确性;另一方面,在研究汽车避碰路径规划问题中,考虑了驾驶意图、汽车运动状态和道路边界约束等因素提出了改进的道路势场模型,进而规划出汽车避碰路径更能适应复杂多变的道路交通环境。本文提出的模型与方法可以为实现车路协同式条件下的汽车智能避碰系统提供理论基础。研究成果对于提高道路交通安全水平,防止汽车碰撞事故的发生,以及促进智能交通系统在我国的发展及实现具有重要意义。
[Abstract]:The frequent occurrence of road traffic accidents seriously affects people's life and property safety and road traffic passing efficiency. The automobile collision accident is one of the most important accident patterns in road traffic accidents, and how to effectively prevent the occurrence of automobile collision accidents and to improve the running safety of road vehicles has always been a common problem faced by many scholars in the field of automobile safety and auxiliary driving. In the actual road traffic environment, the driver's decision-making/ control behavior, the car's motion state, the road traffic environment (road surface wet-slip degree, road line type, etc.) can lead to the occurrence of the automobile collision accident. The domestic and foreign scholars have carried out in-depth research on the automobile intelligent collision avoidance technology, but so far, there are many problems to be solved for the traffic safety state identification method under the complicated road traffic environment and the automobile collision avoidance decision in the dynamic road traffic environment. how to combine that driver's factor with the dynamic road traffic environment to evaluate the effect of the driving intention/ driving behavior, the car's motion state and the vehicle's outside environment on the safety of the vehicle; How to solve the problem of handling and integration of the multi-source and heterogeneous information related to the traffic safety under the cooperative-sensing environment of the 鈥淢an-Vehicle-Road鈥，
本文编号：2358236
[Abstract]:The frequent occurrence of road traffic accidents seriously affects people's life and property safety and road traffic passing efficiency. The automobile collision accident is one of the most important accident patterns in road traffic accidents, and how to effectively prevent the occurrence of automobile collision accidents and to improve the running safety of road vehicles has always been a common problem faced by many scholars in the field of automobile safety and auxiliary driving. In the actual road traffic environment, the driver's decision-making/ control behavior, the car's motion state, the road traffic environment (road surface wet-slip degree, road line type, etc.) can lead to the occurrence of the automobile collision accident. The domestic and foreign scholars have carried out in-depth research on the automobile intelligent collision avoidance technology, but so far, there are many problems to be solved for the traffic safety state identification method under the complicated road traffic environment and the automobile collision avoidance decision in the dynamic road traffic environment. how to combine that driver's factor with the dynamic road traffic environment to evaluate the effect of the driving intention/ driving behavior, the car's motion state and the vehicle's outside environment on the safety of the vehicle; How to solve the problem of handling and integration of the multi-source and heterogeneous information related to the traffic safety under the cooperative-sensing environment of the 鈥淢an-Vehicle-Road鈥，
本文编号：2358236
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2358236.html
