车联网环境下多目标约束的分布式动态路径诱导研究

Traditional route guidance has the problem of that its information source and spread method are single. Vehicular Ad hoc Networks (VANET) changes vehicle from single means of transportation to information platform, which can solve the problem above. This project will research multi-objective constraint distributed dynamic route guidance. Firstly, objective factors and criterion factors of QoS (Quality of Service) of routing in VANET will be analyzed, weight of every criterion factor will be given by Analytic Hierarchy Process (AHP), on the basis, criterion model will be set up. Secondly, in order to satisfy to the needs of vehicle to vehicle communication, the routing algorithm of vehicle to vehicle in VANET will be proposed on the base of Particle Swarm Optimization (PSO) and criterion model, so the route set can be generated. At last, objective factors of selecting route by driver will be investigated and analyzed, objective influence function will be proposed, and PSO will be used again to optimize the best route based on the individual preference and objective environment. The achievement of this project can be applied to traffic jam dissipating, route guidance, and safety warning, and can provide theory evidence for developing and promoting of VANET production.

传统路径诱导存在着信息来源及发布手段单一、诱导效率低的问题，而车联网将车辆从单一的运输工具转变为信息平台，为解决该问题提供了新思路。本项目研究车联网环境下多目标约束的分布式动态路径诱导方法，首先，分析车联网路由服务质量的目标因素及判据因子，利用层次分析法对诸多判据因子进行评价并赋予权值，从而构建路由判据模型；其次，为了适应车辆- - 车辆之间的路由需求，利用粒子群优化算法及路由判据模型确定车辆节点发送信息的下一跳节点，并在此基础上，生成驾驶员出行路径集；最后，调查并解析驾驶员出行路径选择的目标因素，构建综合目标函数，并再次利用粒子群优化算法迭代搜索适合于驾驶员个体偏好及客观环境的最优路径。该项目的研究成果将为车联网环境下的拥堵消散、路径诱导、安全预警等前沿关键问题提供新思路和新方法，为车联网的系列产品的开发和推广提供理论依据。

传统路径诱导系统效率低下的问题，本项目研究车联网环境下多目标约束的分布式动态路径诱导方法，主要成果如下：（1）通过对驾驶员出行过程中的路径选择模式调查，确定了驾驶员路径选择的3种不同模式；（2）对影响驾驶员出行路径选择模式的因素进行筛选，并实现了对其取值范围的分类；（3）利用C4.5决策树分类法实现了对调查数据的训练与测试，实现了在不同影响属性组合下的驾驶员出行路径选择模式分类；（4）根据车辆的位置、速度、邻居节点等相关参数，选举簇首，实现了车联网的车辆自组分簇；（5）通过研究PSO粒子的特性，制定了用于路由及路径诱导的粒子、速度与迭代规则；（6）设计了针对路由算法的PSO适应度函数，构建了融合分簇与PSO的车联网高效路由算法；（7）融合前期调查采用的路径选择模式影响因素，结合出行费用函数等参数，构建了用于路径诱导的PSO适应度函数，实现了多目标约束的分布式动态路径诱导。