车载自组网实时协作定位系统及多数据源融合算法研究

This project focuses on the key technologies for vehicular ad hoc networks (VANETs), which have emerged as a radically new paradigm for the design of networking protocols, mobility models, and a variety of new applications. This technology has its roots in on-board sensors in vehicles, vehicle localization systems, and algorithms or protocols for ad hoc networks, which enable vehicle-to-vehicle communications without infrastructure equipment (e.g., base stations in cellular systems). A fundamental yet open issue is addressed: how rapidly and efficiently can position of vehicles be estimated in a vehicular ad hoc network? To address the unique challenges presented by mobility-induced localization delay problem, a complimentary approach to cooperative localization and its data fusion is employed. The project focuses on three issues: (1) Aiming at topological dynamic and real-time requirements of vehicle localization, to construct analysis method and model of real-time localization, develop new multi-scale real-time evaluation strategies and evaluation indicators. (2) Exploring the impact of the high-speed mobility of vehicles on localization performance, to develop new real-time cooperative localization model and algorithm that integrate characteristics of vehicle movement. (3) Exploring the impact of the network layer protocol on localization real-time performance, to develop multi-channel adaptive spectrum allocation technology that provides network layer support for real-time cooperative localization. (4) Identifying the theoretical framework of cooperative data fusion, to develop real-time data fusion architecture. (5) To develop a simulation and verification platform that supports the practical deployment of vehicular cooperative localization system. As VANETs advance large-scale and social networks, this project addresses an acute and timely demand for exploring fundamental principles of mobility-induced localization delay problem, which have not been studied systematically, but have tremendous impact on vehicular localization system designs, optimization techniques for real-time analysis and estimation, and modeling of vehicular real-time localization.

满足应用环境复杂性要求的车辆实时协作定位技术是车载自组网(VANET: Vehicular ad hoc networks)应用的关键环节。本研究针对车辆运动速度快、定位时延问题突出等实际情况，提出满足实时、可靠要求的车载定位解决方案。首先，针对VANET的拓扑动态特性和定位的实时性要求，研究网络动态性、异构性和连通性等要素之间的制约关系，提出多尺度实时评价策略，研究车载定位的实时性分析方法，建立实时性评估模型。通过挖掘城市环境车辆运动特点，提出具有实时保证的车辆协作定位模型及其实时定位算法。通过分析网络层行为对协作定位的影响，研究多信道动态频谱分配问题，为车载协作定位的实时性提供网络保障。最后，对支持多信息源定位的数据融合机制进行研究，给出实时数据融合模型，据此提出支持容错和实时的融合体系结构，进而构建、优化实时定位系统仿真和物理验证平台，支撑车载定位面向实时、可靠的应用部署。

满足应用环境复杂性要求的车辆实时协作定位技术是车载自组网(VANET:Vehicular ad hoc networks)应用的关键环节。本研究针对车辆运动速度快、定位时延问题突出等实际情况，提出满足实时、可靠要求的车载定位解决方案。首先，针对VANET 的拓扑动态特性和定位的实时性要求，研究网络动态性、异构性和连通性等要素之间的制约关系，提出多尺度实时评价策略，研究车载定位的实时性分析方法，建立实时性评估模型。通过挖掘城市环境车辆运动特点，提出具有实时保证的车辆协作定位模型及其实时定位算法。通过分析网络层行为对协作定位的影响，研究多信道动态频谱分配问题，为车载协作定位的实时性提供网络保障。最后，对支持多信息源定位的数据融合机制进行研究，给出实时数据融合模型，据此提出支持容错和实时的融合体系结构，进而构建、优化实时定位系统仿真和物理验证平台，支撑车载定位面向实时、可靠的应用部署。