基于LBS的交通信息主动推送模型与试验验证

发布时间：2018-11-26 13:23

【摘要】：机动车保有量激增,导致交通供需矛盾日益突出,交通信息服务作为调整交通流时空分布、提高路网通行能力的有效手段,广泛应用于城市交通管理。研究表明现实生活中约80%的信息与位置息息相关,因此将位置服务与交通信息服务结合已经成为交通信息服务的发展趋势。基于LBS的交通信息主动推送模型将有效解决行车环境中信息过载问题,减少出行者时间资源浪费,缓解交通供需矛盾。借助所提出的融合定位算法提供位置信息,本文提出了基于LBS的交通信息主动推送模型,并完成了主动推送系统总体设计及软硬件设计,结合两种主动推送场景验证了主动推送模型的有效性。首先,从信息推送、位置获取等方面总结国内外研究现状,提出了本文的研究内容和技术路线。接着,在介绍位置服务系统的体系结构、关键技术基础上,总结分析了目前国内外交通信息服务系统应用现状及问题。然后,分析现有定位方法的不足,充分利用各个定位方法在时空上互补特性,提出了基于BP神经网络的融合定位算法,利用卡尔曼滤波和量子粒子群实现GPS、DR、WIFI的多源融合定位;借助融合定位算法提供的位置信息,从位置服务角度出发设计了基于推送状态、推送优先级、推送强度以及推送时长等特征参数的主动推送模型。为验证模型有效性构建了主动推送系统,详细给出了系统的硬件、软件、数据库等设计。最后,基于设计的主动推送系统对所提出的融合定位算法和主动推送模型进行了实车验证,考虑推送内容设计事故预警、路况提示两种场景,通过分析主动推送模型特征参数等来验证推送模型有效性。并对主动推送响应时间进行分布特征分析,同时进行了推送距离、行驶速度对响应时间的敏感性分析。利用所设计系统对模型进行分析实验结果表明,相较于其他两种定位算法,所提出的融合定位算法定位精度分别提高87.06%和68.71%,可为主动推送模型提供有效位置信息。所设计的主动推送系统可在1-3s内为车辆提供实时的基于LBS的交通信息服务,并且该系统可随车辆位置、路网熟悉度以及行驶速度实时调整主动信息服务策略;此外,主动推送模型特征参数敏感性分析表明,推送距离、行驶速度对响应时间有影响,响应时间随推送距离增加呈现先迟缓增长后快速增长再稳定的变化趋势,行驶速度也呈现类似的影响规律。随着行驶速度以及推送距离的逐渐增大,响应时间逐渐增大并趋近于3s。
[Abstract]:The increasing number of motor vehicles leads to the increasingly prominent contradiction between supply and demand. As an effective means to adjust the space-time distribution of traffic flow and improve the capacity of road network, traffic information service is widely used in urban traffic management. The research shows that about 80% of the information is closely related to the location in real life, so the combination of location service and traffic information service has become the development trend of traffic information service. The active push model of traffic information based on LBS will effectively solve the problem of information overload in traffic environment, reduce the waste of time and resources of travelers, and alleviate the contradiction between traffic supply and demand. With the help of the proposed fusion location algorithm to provide location information, this paper proposes an active push model of traffic information based on LBS, and completes the overall design of active push system and the design of software and hardware. Combining two kinds of active push scenarios, the validity of the active push model is verified. Firstly, this paper summarizes the status quo of information push and location acquisition at home and abroad, and puts forward the research content and technical route of this paper. Then, on the basis of introducing the architecture and key technologies of the location service system, this paper summarizes and analyzes the current situation and problems of the application of the transportation information service system at home and abroad. Then, by analyzing the shortcomings of the existing localization methods and making full use of the complementary characteristics of each localization method in time and space, a fusion localization algorithm based on BP neural network is proposed. The GPS,DR, is realized by using Kalman filter and quantum particle swarm optimization (QPSO). Multi-source fusion localization of WIFI; Based on the location information provided by the fusion location algorithm, an active push model based on push state, push priority, push intensity and push time is designed from the point of view of location service. In order to verify the validity of the model, the active push system is constructed, and the hardware, software and database design of the system are given in detail. Finally, based on the design of the active push system, the proposed fusion location algorithm and the active push model are verified by the actual vehicle, considering the push content design accident early warning, the road situation hint two kinds of scenarios, The validity of the push model is verified by analyzing the characteristic parameters of the active push model. The distribution characteristics of active push response time and the sensitivity of push distance and travel speed to response time are analyzed. The experimental results show that compared with the other two localization algorithms, the accuracy of the proposed fusion localization algorithm is increased by 87.06% and 68.71%, respectively. It can provide effective location information for active push model. The designed active push system can provide real-time traffic information service based on LBS for vehicles within 1-3s, and the system can adjust the active information service strategy in real time with vehicle location, road network familiarity and driving speed. In addition, the sensitivity analysis of the characteristic parameters of the active push model shows that the push distance and the driving speed have an effect on the response time, and the response time increases slowly first and then steadily with the increase of the push distance. The driving speed also presents the similar influence law. With the increase of driving speed and push distance, the response time increases and approaches to 3 s.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495

【参考文献】