面向预约的状态驱动无人车控制与调度研究

发布时间：2018-07-31 05:36

【摘要】：随着多维感知、智能驾驶、移动通信等技术的发展进步,智能化的无人车辆已逐步成为现实,并在智能交通系统中发挥着重要作用。目前无人车的自主行驶技术日趋成熟,而基于V2V、V2I等通信技术的多车自主协同问题正成为智能交通领域的研究热点。本文依托国家自然科学基金课题、法中博士后科学与应用基金博士后研究课题等,结合未来智能交通系统中的多异构无人车协同场景,研究并提出了面向集中式协作的优先级调度算法,以保证系统中高优先级车辆的通过效率。本文首先对交通系统中的基本对象进行了分析建模,并在分析无人车运动的时空域特征基础上,定义了V2I消息类型和通信机制,建立了无人车动作状态模型和支持消息预约及优先级调度的数学模型;其次,将物理环境与无人车运动特性相结合,设计了面向预约的无人车集中式调度机制;在深入分析高优先级无人车的直接、间接阻塞原理之后,提出了基于时间窗口的全局延迟时间评估方法,进而设计了基于继承的无人车优先级动态管理机制;面向单临界区、多临界区等不同场景研究并设计了基于优先级继承的多无人车调度算法。最后,基于Windows平台开发了模拟仿真软件,实现了论文所研究并设计的模型、机制和算法,以支持对研究工作的实验验证及评估。实验结果表明,本文所设计的交通对象模型可以有效刻画复杂交通环境并支持异构无人车协作机制的研究;所提出的调度算法较现有基于到达时间或队列长度等的调度方法而言,均可有效降低高优先级车辆的延迟时间,满足交通系统对多异构无人车的调度需求。总体上,本文研究工作对智能交通领域的多车协同等相关研究具有实际意义和参考价值。
[Abstract]:With the development of multi-dimension perception, intelligent driving, mobile communication and other technologies, intelligent unmanned vehicles have gradually become a reality and play an important role in the intelligent transportation system. At present, autonomous driving technology of unmanned vehicles is becoming more and more mature, and the problem of multi-vehicle autonomous collaboration based on V2V / V2I and other communication technologies is becoming a research hotspot in the field of intelligent transportation. Based on the project of the National Natural Science Foundation and the postdoctoral research project of the Law and Application Foundation, this paper combines the multi-heterogeneous unmanned vehicle collaboration scene in the future intelligent transportation system. A centralized collaborative priority scheduling algorithm is proposed to ensure the efficiency of high priority vehicles in the system. In this paper, the basic objects in the traffic system are analyzed and modeled, and the V2I message type and communication mechanism are defined on the basis of analyzing the space-time domain characteristics of the motion of the unmanned vehicle. The action state model of UAV and the mathematical model supporting message reservation and priority scheduling are established. Secondly, the centralized scheduling mechanism for UAV is designed by combining the physical environment with the motion characteristics of UAV. After deeply analyzing the direct and indirect blocking principle of high priority unmanned vehicle, a global delay time evaluation method based on time window is proposed, and then a dynamic management mechanism based on inheritance is designed, which is oriented to single critical region. The multi-vehicle scheduling algorithm based on priority inheritance is studied and designed in different scenarios such as multi-critical region. Finally, the simulation software based on Windows platform is developed, and the model, mechanism and algorithm studied and designed in this paper are implemented to support the experimental verification and evaluation of the research work. The experimental results show that the traffic object model designed in this paper can effectively describe the complex traffic environment and support the research of the cooperation mechanism of heterogeneous unmanned vehicles. Compared with the existing scheduling methods based on arrival time or queue length, the proposed scheduling algorithm can effectively reduce the delay time of high-priority vehicles and meet the scheduling requirements of multi-heterogeneous vehicles in traffic systems. In general, the research work in this paper has practical significance and reference value for the research of multi-vehicle coordination in the field of intelligent transportation.
【学位授予单位】：西北工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;U463.6

【参考文献】