基于BP神经网络的碰撞预测及车联网MAC层协议设计

发布时间：2018-02-27 20:07

本文关键词： 主动安全 BP神经网络碰撞预测 D-MAC协议　出处：《西安电子科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：近年来,交通事故呈不断上升的趋势,尤其是高速公路上的追尾事故频发,因此人们越来越多的关注如何保障高速公路行车安全。目前保障汽车行驶安全的技术主要分为被动安全技术和主动安全技术。主动安全技术能够根据当前车辆的运动状态和周围环境信息,对潜在的冲突予以判断。主动安全技术能从源头上抑制交通事故的发生,所以对汽车主动安全技术的研究具有重要的意义。车辆状态预测是汽车主动安全技术的基础,根据以往车辆碰撞预测的成果,本文针对车辆碰撞概率的预测展开研究。现有的车辆碰撞概率计算方法只是结合运动学公式和车辆的分布情况对车辆追尾碰撞概率进行计算,没有考虑车辆追尾碰撞的真实场景,预测结果有偏差。为了提高车辆碰撞概率预测的准确度,本文综合考虑造成车辆碰撞的驾驶员、车辆、道路和环境等因素,采用BP神经网络的方法对车辆的追尾碰撞情况进行预测。由于BP神经网络的初始化连接权值和阈值的选择具有很大的随机性,可能使BP神经网络训练的结果陷入局部最优,而遗传算法具有全局寻优的能力,因此本文选用遗传算法对BP神经网络的初始化连接权值和阈值进行优化,此外,为了改进BP神经网络的收敛速度,本文进一步对BP神经网络的学习速率进行改进。最后利用MATLAB仿真工具对本文中的车辆状态预测算法进行性能验证,结果表明,本文提出的算法能较准确的预测车辆下一时刻的碰撞概率。准确预测到车辆发生碰撞的信息后,确保告警信息可靠及时的发送是十分必要的。为了保障安全信息的可靠及时的传输,同时提高周期性beacon消息接入信道的公平性。本文针对高速公路场景的车辆协同防撞应用,根据以RSU为中心的车队中车辆节点的数目和车队中发生潜在碰撞的车辆节点的数目,设计了保障告警信息可靠及时传输的D-MAC(Dynamic slot Media Access Control,D-MAC)协议。该协议是基于动态TDMA机制,根据当前车队中的实时交通动态确定每帧的时隙数目,并且该协议优先为告警信息分配时隙,同时尽可能地为周期性beacon消息预留传输时隙。最后,在不同的车辆节点密度的情况下,对本文提出的动态介质访问控制协议的性能进行评估,与IEEE802.11p协议相比,D-MAC协议的投递率增加了大约39%,在节点密度较大时,和IEEE802.11p协议相比,D-MAC协议的数据传输时延更低,并且D-MAC协议获取无线信道服务的公平性提高了31%。
[Abstract]:In recent years, traffic accidents have been on the rise, especially the frequent rear-end accidents on highways. Therefore, people are paying more and more attention to how to ensure the safety of motorway. At present, the technology to ensure the safety of vehicle driving is mainly divided into passive safety technology and active safety technology. Active safety technology can be based on the current vehicle. State of motion and surrounding environment information, The active safety technology can restrain the occurrence of the traffic accident from the source, so the research on the active safety technology of the automobile is of great significance. The vehicle state prediction is the basis of the active safety technology of the vehicle. According to the results of vehicle collision prediction in the past, this paper studies the prediction of vehicle collision probability. The existing methods of calculating vehicle collision probability are only based on kinematics formula and vehicle distribution to calculate the rear-end collision probability. In order to improve the accuracy of the prediction of vehicle collision probability, the driver, vehicle, road and environment of vehicle collision are considered comprehensively in this paper. The method of BP neural network is used to predict the rear-end collision of the vehicle. Because of the randomness of the selection of the initial connection weight and the threshold value of BP neural network, the result of BP neural network training may fall into local optimum. The genetic algorithm has the ability of global optimization, so the genetic algorithm is used to optimize the initial connection weight and threshold of BP neural network. In addition, in order to improve the convergence speed of BP neural network, In this paper, the learning rate of BP neural network is further improved. Finally, the performance of the vehicle state prediction algorithm in this paper is verified by using MATLAB simulation tool, and the results show that, The algorithm proposed in this paper can accurately predict the collision probability of vehicles at the next moment. It is very necessary to ensure the reliable and timely transmission of alarm information. In order to ensure the reliable and timely transmission of security information and to improve the fairness of periodic beacon message access channel, this paper aims at the vehicle anti-collision application of freeway scene. According to the number of vehicle nodes in the vehicle fleet centered on RSU and the number of vehicle nodes with potential collision in the vehicle fleet, a D-MAC dynamic slot Media Access Control D-MACCprotocol is designed to guarantee the reliable and timely transmission of alarm information. The protocol is based on the dynamic TDMA mechanism. The number of time slots per frame is determined according to the real-time traffic dynamics in the current motorcade, and the protocol preferentially allocates time slots for alarm information, while reserving transmission slots for periodic beacon messages as far as possible. Finally, In the case of different vehicle node density, the performance of the proposed dynamic media access control protocol is evaluated. Compared with the IEEE802.11p protocol, the delivery rate of the D-MAC protocol increases by about 39 percent. Compared with the IEEE802.11p protocol, the data transmission delay of the D-MAC protocol is lower, and the fairness of the D-MAC protocol to obtain wireless channel services is improved by 31.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP183

【参考文献】