中文摘要：

为了克服前期提出的汽车自组织网络（VANETs）链路时延预测（LDP）模型只能预测高速公路直行场景的链路时延的局限，扩展了LDP模型的功能，构建了扩展的LDP模型——ELDP模型，并对该模型的链路时延预测过程进行了理论分析，对其预测性能进行了仿真分析。ELDP模型考虑了两车相对速度值的分布、相对速度驶过距离的变化以及在交叉口转弯对两车之间链路时延的影响，可以动态地预测车辆转弯时的两车链路时延，能够估计任意一对汽车预期要发生的链路时延。该模型实现取决于对相对速度值的分布参数估计方法、基于指数移动平均法对相对速度值的异常处理以及对在道路交叉口汽车转弯角度对链路时延的影响的估计。仿真结果表明，ELDP模型能够非常准确地预测高速公路中两车的链路时延。

英文摘要：

The link duration prediction （LDP） model proposed in the earlier stage of the study for vehicular ad hoc networks （VANETs） was functionally extended, and an extended LDP （ELDP） model was formed to break the LDP model＇ s functional limitation of only working in the straight highway environment. The LDP process of the ELDP model was theoretically analyzed, and the model＇ s predicting performance was investigated by simulation. With the impacts of two vehicles＇ relative velocity, relative distance and turning angles at an intersection on link duration considered, The ELDP model can dynamically predict the link duration between two vehicles when they making a turn, and estimate the expected link duration between any pair of connected vehicles. The model＇ s implementation depends on the approach for parameters estimation of relative velocity value distribution, exponential moving average （EMA） method processes relative speed value exception and the estimation of turning angle＇ s impacts on link duration. The simulation results show that the ELDP model can very accurately predict link duration.